{ "generated_at": "2026-10-05", "count": 3302, "entries": [ { "id": "xtts-v2", "name": "XTTS-v2", "type": "tts", "country": "INTL", "org": "Coqui", "license": "coqui-public-model-license", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/coqui/XTTS-v2" }, "notes": "Coqui's multilingual voice cloning, Arabic support, 6s cloning", "metrics": { "downloads": 6665636, "likes": 3846, "lastModified": "2023-12-11" } }, { "id": "whisper-large-v3-turbo", "name": "whisper-large-v3-turbo", "type": "asr", "country": "INTL", "org": "OpenAI", "license": "mit", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/openai/whisper-large-v3-turbo" }, "notes": "4x faster Whisper, Arabic support, MIT license", "base_model": [ "openai/whisper-large-v3" ], "metrics": { "downloads": 6295038, "likes": 3412, "lastModified": "2024-10-04" } }, { "id": "openai-whisper-large-v3", "name": "openai/whisper-large-v3", "type": "asr", "country": "INTL", "org": "OpenAI", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/openai/whisper-large-v3" }, "notes": "Supports Arabic among many languages", "metrics": { "downloads": 4050188, "likes": 6561, "lastModified": "2024-08-12" } }, { "id": "multilingual-chatterbox", "name": "Multilingual Chatterbox", "type": "tts", "country": "INTL", "org": "Resemble AI", "license": "mit", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/ResembleAI/chatterbox" }, "year": 2025, "notes": "Resemble AI Chatterbox multilingual TTS model covering 22 languages including Arabic.", "metrics": { "downloads": 1668253, "likes": 1823, "lastModified": "2026-06-10" } }, { "id": "wav2vec2-large-xlsr-53-arabic", "name": "wav2vec2-large-xlsr-53-arabic", "type": "asr", "country": "INTL", "org": "jonatasgrosman", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/jonatasgrosman/wav2vec2-large-xlsr-53-arabic" }, "notes": "Fine-tuned on Common Voice & Arabic Speech Corpus", "metrics": { "downloads": 1514391, "likes": 55, "lastModified": "2022-12-14" } }, { "id": "arabertv02", "name": "AraBERTv02", "type": "llm", "country": "LB", "org": "AUB MIND Lab", "license": "unknown", "modality": "text", "tasks": [ "encoder" ], "links": { "hf": "https://huggingface.co/aubmindlab/bert-base-arabertv02" }, "notes": "Improved tokenization (135M params, 12M+ downloads)", "metrics": { "downloads": 557389, "likes": 51, "lastModified": "2024-03-26" } }, { "id": "phi-4", "name": "Phi-4", "type": "llm", "country": "INTL", "org": "Microsoft", "license": "mit", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/microsoft/phi-4" }, "size": "4B", "notes": "Multilingual with Arabic, compact & efficient", "metrics": { "downloads": 449588, "likes": 2316, "lastModified": "2026-07-14" } }, { "id": "gemma4-e4b-claims-comparison", "name": "gemma4 e4b claims comparison", "type": "llm", "country": "INTL", "org": "k-chirkunov", "license": "gemma", "modality": "text", "tasks": [ "nli", "claims-comparison" ], "links": { "hf": "https://huggingface.co/k-chirkunov/gemma4-e4b-claims-comparison" }, "year": 2026, "notes": "Gemma 4 E4B LoRA fine-tune for Arabic dialectal claims comparison (NLI).", "base_model": [ "google/gemma-4-e4b-it" ], "metrics": { "downloads": 437351, "likes": 2, "lastModified": "2026-08-05" } }, { "id": "llama-3-3", "name": "Llama 3.3", "type": "llm", "country": "INTL", "org": "Meta", "license": "llama3.3", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/meta-llama/Llama-3.3-70B-Instruct" }, "size": "70B", "notes": "Strong Arabic performance", "base_model": [ "meta-llama/llama-3.1-70b" ], "metrics": { "downloads": 399695, "likes": 3099, "lastModified": "2024-12-21" } }, { "id": "mms-1b-all", "name": "MMS-1b-all", "type": "asr", "country": "INTL", "org": "Meta", "license": "cc-by-nc-4.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/facebook/mms-1b-all" }, "notes": "Meta's Massively Multilingual Speech, ASR for 1100+ languages", "metrics": { "downloads": 292617, "likes": 207, "lastModified": "2023-06-15" } }, { "id": "seamlessm4t-v2", "name": "SeamlessM4T v2", "type": "asr", "country": "INTL", "org": "Meta", "license": "cc-by-nc-4.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/facebook/seamless-m4t-v2-large" }, "notes": "Meta's all-in-one ASR + translation, ~100 languages inc. Arabic", "metrics": { "downloads": 286661, "likes": 1024, "lastModified": "2024-01-04" } }, { "id": "prophet-mosque-library", "name": "prophet mosque library", "type": "dataset", "country": "INTL", "org": "ieasybooks", "license": "mit", "modality": "text", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/datasets/ieasybooks-org/prophet-mosque-library" }, "size": "10K–100K rows", "year": 2025, "notes": "Prophet’s Mosque Library is one of the primary resources for Islamic books.", "metrics": { "downloads": 284301, "likes": 6, "lastModified": "2025-05-14" } }, { "id": "fanar-2-27b-instruct", "name": "Fanar-2-27B-Instruct", "type": "llm", "country": "QA", "org": "QCRI", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/QCRI/Fanar-2-27B-Instruct", "paper": "https://arxiv.org/abs/2603.16397" }, "base_model": [ "google/gemma-3-27b-pt" ], "size": "27B", "year": 2026, "notes": "Fanar 2 Arabic-English instruct model built on Gemma 3 27B, released March 2026.", "metrics": { "downloads": 204831, "likes": 16, "lastModified": "2026-03-25" } }, { "id": "voxtral-mini", "name": "Voxtral Mini", "type": "asr", "country": "INTL", "org": "Mistral AI", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/mistralai/Voxtral-Mini-3B-2507" }, "notes": "Mistral's speech model, 3B, Arabic support, Apache 2.0", "metrics": { "downloads": 175023, "likes": 675, "lastModified": "2025-07-28" } }, { "id": "opus-mt-ar-en", "name": "opus-mt-ar-en", "type": "llm", "country": "INTL", "org": "Helsinki-NLP", "license": "apache-2.0", "modality": "text", "tasks": [ "task-specific", "translation" ], "links": { "hf": "https://huggingface.co/Helsinki-NLP/opus-mt-ar-en" }, "notes": "Translation AR→EN - Helsinki-NLP (12.4M+ downloads)", "metrics": { "downloads": 132239, "likes": 50, "lastModified": "2023-08-16" } }, { "id": "waqfeya-library", "name": "Waqfeya Library", "type": "dataset", "country": "INTL", "org": "ieasybooks", "license": "mit", "modality": "text", "tasks": [ "pretraining", "ocr" ], "links": { "hf": "https://huggingface.co/datasets/ieasybooks-org/waqfeya-library" }, "size": "10K-100K books", "year": 2025, "dialects": [ "classical", "msa" ], "notes": "Book texts and PDFs from the Waqfeya Islamic library, 10K-100K files.", "metrics": { "downloads": 129453, "likes": 12, "lastModified": "2025-05-14" } }, { "id": "arapoembert", "name": "AraPoemBERT", "type": "llm", "country": "SA", "org": "King Saud University", "license": "cc-by-nc-4.0", "modality": "text", "tasks": [ "classification" ], "links": { "hf": "https://huggingface.co/faisalq/bert-base-arapoembert" }, "year": 2024, "notes": "BERT model pretrained on Arabic poetry for meter, rhyme, sentiment classification.", "metrics": { "downloads": 109004, "likes": 2, "lastModified": "2024-05-23" } }, { "id": "bert-base-arabic-camelbert-mix-sentiment", "name": "bert-base-arabic-camelbert-mix-sentiment", "type": "llm", "country": "AE", "org": "CAMeL Lab, NYUAD", "license": "apache-2.0", "modality": "text", "tasks": [ "sentiment" ], "links": { "hf": "https://huggingface.co/CAMeL-Lab/bert-base-arabic-camelbert-mix-sentiment" }, "year": 2022, "dialects": [ "mixed" ], "notes": "CAMeLBERT-Mix fine-tuned for Arabic sentiment analysis.", "metrics": { "downloads": 102574, "likes": 8, "lastModified": "2021-10-17" } }, { "id": "marbertv2", "name": "MARBERTv2", "type": "llm", "country": "INTL", "org": "UBC-NLP", "license": "unknown", "modality": "text", "tasks": [ "encoder" ], "links": { "hf": "https://huggingface.co/UBC-NLP/MARBERTv2" }, "dialects": [ "mixed" ], "notes": "Updated with improved dialectal coverage", "metrics": { "downloads": 102274, "likes": 14, "lastModified": "2022-03-30" } }, { "id": "sentimentareng", "name": "SentimentArEng", "type": "llm", "country": "INTL", "org": "qandos0", "license": "unknown", "modality": "text", "tasks": [ "encoder", "sentiment" ], "links": { "hf": "https://huggingface.co/qandos0/SentimentArEng" }, "year": 2023, "notes": "Fine-tune of cardiffnlp twitter-xlm-roberta-base-sentiment for Arabic-English sentiment classification.", "base_model": [ "cardiffnlp/twitter-xlm-roberta-base-sentiment" ], "metrics": { "downloads": 96525, "likes": 0, "lastModified": "2023-12-14" } }, { "id": "shamela-waqfeya-library", "name": "shamela waqfeya library", "type": "dataset", "country": "INTL", "org": "ieasybooks", "license": "mit", "modality": "text", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/datasets/ieasybooks-org/shamela-waqfeya-library" }, "size": "1K–10K rows", "year": 2025, "notes": "Shamela Waqfeya is one of the primary online resources for Islamic books, similar to Shamela.", "metrics": { "downloads": 90500, "likes": 4, "lastModified": "2025-05-14" } }, { "id": "sard", "name": "SARD", "type": "dataset", "country": "SA", "org": "riotu-lab", "license": "cc-by-nc-nd-4.0", "modality": "vision", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/datasets/riotu-lab/SARD" }, "notes": "Synthetic Arabic OCR dataset for recognition training", "metrics": { "downloads": 56805, "likes": 14, "lastModified": "2026-05-20" } }, { "id": "bert-base-arabic-camelbert-da-sentiment", "name": "bert-base-arabic-camelbert-da-sentiment", "type": "llm", "country": "AE", "org": "CAMeL Lab, NYUAD", "license": "apache-2.0", "modality": "text", "tasks": [ "sentiment" ], "links": { "hf": "https://huggingface.co/CAMeL-Lab/bert-base-arabic-camelbert-da-sentiment" }, "year": 2022, "dialects": [ "mixed" ], "notes": "CAMeLBERT-DA fine-tuned for Arabic sentiment analysis.", "metrics": { "downloads": 51077, "likes": 51, "lastModified": "2021-10-17" } }, { "id": "cohere-transcribe-arabic-07-2026", "name": "cohere-transcribe-arabic-07-2026", "type": "asr", "country": "INTL", "org": "Cohere Labs", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/CohereLabs/cohere-transcribe-arabic-07-2026" }, "size": "2.1B", "year": 2026, "dialects": [ "msa" ], "notes": "Cohere Transcribe specialised for Arabic speech recognition.", "base_model": [ "coherelabs/cohere-transcribe-03-2026" ], "metrics": { "downloads": 47128, "likes": 212, "lastModified": "2026-07-13" } }, { "id": "opus-mt-tc-big-ar-en", "name": "opus mt tc big ar en", "type": "llm", "country": "INTL", "org": "Helsinki-NLP", "license": "cc-by-4.0", "modality": "text", "tasks": [ "translation" ], "links": { "hf": "https://huggingface.co/Helsinki-NLP/opus-mt-tc-big-ar-en" }, "year": 2023, "notes": "Neural machine translation model for translating from Arabic (ar) to English (en).", "metrics": { "downloads": 41224, "likes": 22, "lastModified": "2023-08-16" } }, { "id": "arabic-books", "name": "Arabic Books", "type": "dataset", "country": "INTL", "org": "MohamedRashad", "license": "gpl-3.0", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/datasets/MohamedRashad/arabic-books" }, "size": "8.5k books", "year": 2024, "dialects": [ "msa" ], "notes": "8,500 rows of full Arabic book texts.", "metrics": { "downloads": 33751, "likes": 3, "lastModified": "2024-11-28" } }, { "id": "xnli", "name": "XNLI", "type": "dataset", "country": "INTL", "org": "Facebook AI Research", "license": "cc-by-nc-4.0", "modality": "text", "tasks": [ "nli" ], "links": { "github": "https://github.com/facebookresearch/XNLI", "hf": "https://huggingface.co/datasets/facebook/xnli", "paper": "https://arxiv.org/pdf/1809.05053.pdf" }, "dialects": [ "msa" ], "size": "7,500 sentences", "year": 2018, "tags": [ "multilingual" ], "notes": "Evaluation set for XLU by extending the development and test sets of the Multi-Genre Natural Language Inference Corpus (MultiNLI) to 15 languages,", "metrics": { "downloads": 31749, "likes": 73, "lastModified": "2024-01-05" } }, { "id": "shamela4-full-db", "name": "Shamela4 Full DB", "type": "dataset", "country": "INTL", "org": "AuthenticIlm", "license": "mit", "modality": "text", "tasks": [ "ner", "qa" ], "links": { "hf": "https://huggingface.co/datasets/AuthenticIlm/Shamela4_Full_DB" }, "size": "10M–100M rows", "year": 2026, "notes": "A complete extraction of al-Maktaba al-Shamela (الشاملة) v4, containing 8,589 books across 40 categories of classical Islamic sciences.", "metrics": { "downloads": 30637, "likes": 30, "lastModified": "2026-05-19" } }, { "id": "wikiann", "name": "wikiann", "type": "dataset", "country": "INTL", "org": "Rensselaer Polytechnic Institute", "license": "unknown", "modality": "text", "tasks": [ "ner" ], "links": { "hf": "https://huggingface.co/datasets/unimelb-nlp/wikiann", "paper": "https://aclanthology.org/P17-1178.pdf" }, "dialects": [ "msa" ], "size": "185,000 tokens", "year": 2017, "tags": [ "multilingual" ], "notes": "Name tagging and linking for 282 languages from Wikipedia, with an Arabic subset.", "metrics": { "downloads": 30175, "likes": 125, "lastModified": "2024-02-22" } }, { "id": "opus-mt-en-ar", "name": "opus-mt-en-ar", "type": "llm", "country": "INTL", "org": "Helsinki-NLP", "license": "apache-2.0", "modality": "text", "tasks": [ "task-specific", "translation" ], "links": { "hf": "https://huggingface.co/Helsinki-NLP/opus-mt-en-ar" }, "notes": "Translation EN→AR - Helsinki-NLP (3.5M+ downloads)", "metrics": { "downloads": 28001, "likes": 48, "lastModified": "2023-08-16" } }, { "id": "whisper-large-v3-turbo-ar-quran", "name": "whisper-large-v3-turbo-ar-quran", "type": "asr", "country": "INTL", "org": "Naazim", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/naazimsnh02/whisper-large-v3-turbo-ar-quran" }, "size": "809M", "year": 2025, "dialects": [ "classical" ], "notes": "Whisper large-v3-turbo fine-tuned on Quran recitation.", "base_model": [ "openai/whisper-large-v3-turbo" ], "metrics": { "downloads": 26512, "likes": 2, "lastModified": "2025-12-08" } }, { "id": "global-mmlu", "name": "Global-MMLU", "type": "benchmark", "country": "INTL", "org": "Cohere For AI", "license": "apache-2.0", "modality": "text", "tasks": [ "evaluation", "qa" ], "links": { "hf": "https://huggingface.co/datasets/CohereForAI/Global-MMLU", "paper": "https://doi.org/10.18653/v1/2025.acl-long.919" }, "dialects": [ "msa" ], "size": "14,285 sentences", "year": 2025, "tags": [ "multilingual" ], "notes": "Multilingual MMLU-style benchmark of 42 languages (Arabic included) with culturally sensitive and culturally agnostic subsets.", "metrics": { "downloads": 25560, "likes": 163, "lastModified": "2025-08-14" } }, { "id": "flores-101", "name": "FLORES-101", "type": "benchmark", "country": "INTL", "org": "Facebook AI Research", "license": "cc-by-sa-4.0", "modality": "text", "tasks": [ "evaluation", "translation" ], "links": { "github": "https://github.com/facebookresearch/flores", "hf": "https://huggingface.co/datasets/gsarti/flores_101", "paper": "https://arxiv.org/pdf/2106.03193.pdf" }, "dialects": [ "msa" ], "size": "3,100,000 tokens", "year": 2021, "tags": [ "multilingual" ], "notes": "Low-resource machine translation benchmark of 101 languages, with Arabic as one of them.", "metrics": { "downloads": 24985, "likes": 33, "lastModified": "2022-10-27" } }, { "id": "qwen3-asr-arabic-uae", "name": "qwen3-asr-arabic-uae", "type": "asr", "country": "AE", "org": "Vadim Belsky", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/vadimbelsky/qwen3-asr-arabic-uae" }, "size": "2B", "year": 2026, "dialects": [ "gulf" ], "notes": "Qwen3-ASR fine-tuned for Emirati Arabic.", "base_model": [ "qwen/qwen3-asr-1.7b" ], "metrics": { "downloads": 23661, "likes": 1, "lastModified": "2026-04-04" } }, { "id": "whisper-quran", "name": "Whisper Quran", "type": "asr", "country": "INTL", "org": "Tarteel AI", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/tarteel-ai/whisper-base-ar-quran" }, "dialects": [ "classical" ], "notes": "Whisper fine-tuned for Quranic recitation recognition", "metrics": { "downloads": 23470, "likes": 185, "lastModified": "2022-12-13" } }, { "id": "aya-dataset", "name": "Aya Dataset", "type": "dataset", "country": "INTL", "org": "Cohere For AI Community", "license": "apache-2.0", "modality": "text", "tasks": [ "qa", "instruction-tuning" ], "links": { "hf": "https://huggingface.co/datasets/CohereForAI/aya_dataset", "paper": "https://doi.org/10.18653/v1/2024.acl-long.620" }, "dialects": [ "mixed" ], "size": "14,250 sentences", "year": 2024, "tags": [ "multilingual" ], "notes": "The Aya Dataset is a multilingual instruction fine-tuning dataset curated by an open-science community via Aya Annotation Platform from Cohere For AI.", "metrics": { "downloads": 23214, "likes": 370, "lastModified": "2025-04-15" } }, { "id": "arabic-sbert-100k", "name": "Arabic SBERT 100K", "type": "embedding", "country": "INTL", "org": "akhooli", "license": "unknown", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/akhooli/Arabic-SBERT-100K" }, "year": 2024, "notes": "This is a sentence-transformers model finetuned from aubmindlab/bert-base-arabertv02.", "base_model": [ "aubmindlab/bert-base-arabertv02" ], "metrics": { "downloads": 17433, "likes": 17, "lastModified": "2024-07-27" } }, { "id": "bert-base-arabic-camelbert-msa-ner", "name": "bert-base-arabic-camelbert-msa-ner", "type": "llm", "country": "AE", "org": "CAMeL Lab, NYUAD", "license": "apache-2.0", "modality": "text", "tasks": [ "ner" ], "links": { "hf": "https://huggingface.co/CAMeL-Lab/bert-base-arabic-camelbert-msa-ner" }, "year": 2022, "dialects": [ "msa" ], "notes": "CAMeLBERT-MSA fine-tuned for Arabic named entity recognition.", "metrics": { "downloads": 17108, "likes": 8, "lastModified": "2021-10-17" } }, { "id": "global-mmlu-lite", "name": "Global-MMLU Lite", "type": "dataset", "country": "INTL", "org": "Cohere For AI", "license": "apache-2.0", "modality": "text", "tasks": [ "qa" ], "links": { "hf": "https://huggingface.co/datasets/CohereLabs/Global-MMLU-Lite", "paper": "https://doi.org/10.18653/v1/2025.acl-long.919" }, "dialects": [ "msa" ], "size": "685 sentences", "year": 2025, "tags": [ "multilingual" ], "notes": "Lite version of Global-MMLU with 200 culturally sensitive and 200 culturally agnostic samples per language across 16 languages, including Arabic.", "metrics": { "downloads": 15223, "likes": 43, "lastModified": "2026-06-08" } }, { "id": "gate-arabert-v1", "name": "GATE-AraBert-v1", "type": "embedding", "country": "SA", "org": "Omartificial-Intelligence-Space", "license": "apache-2.0", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/Omartificial-Intelligence-Space/GATE-AraBert-v1" }, "notes": "SOTA on MTEB Arabic STS", "base_model": [ "omartificial-intelligence-space/arabic-triplet-matryoshka-v2" ], "metrics": { "downloads": 14758, "likes": 20, "lastModified": "2025-09-07" } }, { "id": "muaalem-model-v3-2", "name": "muaalem model v3 2", "type": "asr", "country": "INTL", "org": "obadx", "license": "mit", "modality": "speech", "tasks": [ "asr", "quran" ], "links": { "hf": "https://huggingface.co/obadx/muaalem-model-v3_2" }, "dialects": [ "classical" ], "year": 2025, "tags": [ "variants:2" ], "notes": "Arabic automatic speech recognition model fine-tuned from facebook/w2v-bert-2.0 trained on obadx/muaalem-annotated-v3.", "base_model": [ "facebook/w2v-bert-2.0" ], "metrics": { "downloads": 14381, "likes": 11, "lastModified": "2025-09-04" } }, { "id": "quranic-recitation-data", "name": "Quranic Recitation Data", "type": "dataset", "country": "INTL", "org": "zaibihassan", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/datasets/zaibihassan/Quranic-Recitation-Data" }, "size": "100K-1M clips", "year": 2026, "dialects": [ "classical" ], "notes": "Large Quran recitation audio dataset with 100K-1M clips.", "metrics": { "downloads": 13953, "likes": 6, "lastModified": "2026-10-05" } }, { "id": "arabic-common-voice", "name": "Arabic Common Voice", "type": "dataset", "country": "INTL", "org": "Mozilla", "license": "cc0-1.0", "modality": "speech", "tasks": [ "asr" ], "links": { "website": "https://datacollective.mozillafoundation.org/datasets?q=common+voice", "hf": "https://huggingface.co/datasets/legacy-datasets/common_voice", "paper": "https://arxiv.org/pdf/1912.06670.pdf" }, "dialects": [ "msa" ], "size": "85 hours", "year": 2020, "notes": "An open source, multi-language dataset of voices that anyone can use to train speech-enabled applications.", "metrics": { "downloads": 13022, "likes": 148, "lastModified": "2024-08-22" } }, { "id": "f5tts-algerian-darja", "name": "f5tts-algerian-darja", "type": "tts", "country": "DZ", "org": "Touati Kamel", "license": "apache-2.0", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/touati-kamel/f5tts-algerian-darja" }, "year": 2026, "dialects": [ "magh" ], "notes": "F5-TTS fine-tuned on Algerian Darja; Algeria noted.", "base_model": [ "ibrahimsalah/arabic-f5-tts-v2" ], "metrics": { "downloads": 12541, "likes": 0, "lastModified": "2026-09-23" } }, { "id": "asafaya-bert-base-arabic", "name": "asafaya/bert-base-arabic", "type": "embedding", "country": "INTL", "org": "asafaya", "license": "unknown", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/asafaya/bert-base-arabic" }, "notes": "BERT-based Arabic embeddings", "metrics": { "downloads": 12443, "likes": 40, "lastModified": "2023-03-17" } }, { "id": "bert-base-arabic-camelbert-mix-ner", "name": "bert-base-arabic-camelbert-mix-ner", "type": "llm", "country": "AE", "org": "CAMeL Lab, NYUAD", "license": "apache-2.0", "modality": "text", "tasks": [ "ner" ], "links": { "hf": "https://huggingface.co/CAMeL-Lab/bert-base-arabic-camelbert-mix-ner" }, "year": 2022, "dialects": [ "mixed" ], "notes": "CAMeLBERT-Mix fine-tuned for Arabic named entity recognition.", "metrics": { "downloads": 12165, "likes": 15, "lastModified": "2021-10-17" } }, { "id": "bert-base-arabertv2", "name": "bert-base-arabertv2", "type": "llm", "country": "LB", "org": "AUB MIND Lab", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/aubmindlab/bert-base-arabertv2" }, "size": "136M", "year": 2022, "on_device": true, "dialects": [ "msa" ], "notes": "AraBERT v2 base with Farasa-segmented input.", "metrics": { "downloads": 12118, "likes": 46, "lastModified": "2023-08-03" } }, { "id": "fanar-1-9b", "name": "Fanar-1-9B", "type": "llm", "country": "QA", "org": "QCRI", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/QCRI/Fanar-1-9B-Instruct" }, "base_model": [ "google/gemma-2-9b" ], "size": "9B", "dialects": [ "msa" ], "notes": "Arabic-English LLM", "metrics": { "downloads": 11929, "likes": 36, "lastModified": "2025-07-15" } }, { "id": "falcon-arabic", "name": "Falcon Arabic", "type": "llm", "country": "AE", "org": "TII (UAE)", "license": "falcon-llm-license", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/tiiuae/Falcon3-7B-Instruct" }, "base_model": [ "tiiuae/falcon3-7b-base" ], "size": "7B", "notes": "First Arabic model in Falcon series, top of OALL", "metrics": { "downloads": 11713, "likes": 80, "lastModified": "2025-05-31" } }, { "id": "allam", "name": "ALLaM", "type": "llm", "country": "SA", "org": "SDAIA & IBM", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/ALLaM-AI/ALLaM-7B-Instruct-preview" }, "base_model": [ "from-scratch" ], "size": "7B", "dialects": [ "msa" ], "year": 2025, "notes": "Saudi's sovereign model, enterprise-focused", "metrics": { "downloads": 11639, "likes": 195, "lastModified": "2025-07-14" } }, { "id": "allam-7b-instruct-preview", "name": "ALLaM-7B-Instruct-preview", "type": "llm", "country": "SA", "org": "HUMAIN", "license": "apache-2.0", "modality": "text", "tasks": [ "chat", "pretraining" ], "links": { "hf": "https://huggingface.co/humain-ai/ALLaM-7B-Instruct-preview" }, "base_model": [ "from-scratch" ], "dialects": [ "gulf" ], "size": "7B", "on_device": false, "year": 2025, "notes": "ALLaM is a series of powerful language models designed to advance Arabic Language Technology.", "metrics": { "downloads": 11639, "likes": 195, "lastModified": "2025-07-14" } }, { "id": "coda-llm-data", "name": "coda llm data", "type": "dataset", "country": "INTL", "org": "mohameddalii", "license": "apache-2.0", "modality": "speech", "tasks": [ "evaluation", "speech" ], "links": { "hf": "https://huggingface.co/datasets/mohameddalii/coda-llm-data" }, "dialects": [ "gulf" ], "size": "1K–10K rows", "year": 2026, "notes": "This repository contains the full end-to-end dataset, fine-tuning scripts, evaluation suites, load testing harness, and proxy architecture for Coda LLM.", "metrics": { "downloads": 10702, "likes": 0, "lastModified": "2026-10-05" } }, { "id": "mmmlu", "name": "MMMLU", "type": "dataset", "country": "INTL", "org": "OpenAI", "license": "mit", "modality": "text", "tasks": [ "qa" ], "links": { "hf": "https://huggingface.co/datasets/openai/MMMLU" }, "dialects": [ "msa" ], "size": "14,000 sentences", "year": 2024, "tags": [ "multilingual" ], "notes": "Translated the MMLU’s test set into 14 languages using professional human translators.", "metrics": { "downloads": 10489, "likes": 526, "lastModified": "2024-10-16" } }, { "id": "xstorycloze", "name": "XStoryCloze", "type": "dataset", "country": "INTL", "org": "MetaAI", "license": "cc-by-sa-4.0", "modality": "text", "tasks": [ "commonsense" ], "links": { "hf": "https://huggingface.co/datasets/juletxara/xstory_cloze", "paper": "https://doi.org/10.18653/v1/2022.emnlp-main.616" }, "dialects": [ "msa" ], "size": "1,870 sentences", "year": 2022, "tags": [ "multilingual" ], "notes": "XStoryCloze consists of the professionally translated version of the English StoryCloze dataset (Spring 2016 version) to 10 non-English languages.", "metrics": { "downloads": 8986, "likes": 16, "lastModified": "2025-07-23" } }, { "id": "massive", "name": "MASSIVE", "type": "dataset", "country": "INTL", "org": "Amazon", "license": "cc-by-4.0", "modality": "text", "tasks": [ "intent-detection", "slot-filling" ], "links": { "hf": "https://huggingface.co/datasets/qanastek/MASSIVE", "paper": "https://doi.org/10.18653/v1/2023.acl-long.235" }, "dialects": [ "msa" ], "size": "19,521 sentences", "year": 2022, "tags": [ "multilingual" ], "notes": "1M parallel labeled virtual-assistant utterances in 51 languages, with Arabic as a subset.", "metrics": { "downloads": 8972, "likes": 28, "lastModified": "2022-12-23" } }, { "id": "bert-base-arabic-camelbert-mix", "name": "bert-base-arabic-camelbert-mix", "type": "llm", "country": "AE", "org": "CAMeL Lab, NYUAD", "license": "apache-2.0", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/CAMeL-Lab/bert-base-arabic-camelbert-mix" }, "year": 2022, "dialects": [ "mixed" ], "notes": "CAMeLBERT-Mix pretrained on MSA, dialectal and Classical Arabic.", "metrics": { "downloads": 8704, "likes": 19, "lastModified": "2021-09-14" } }, { "id": "wav2vec2-quran-phonetics", "name": "wav2vec2 quran phonetics", "type": "asr", "country": "INTL", "org": "TBOGamer22", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr", "quran" ], "links": { "hf": "https://huggingface.co/TBOGamer22/wav2vec2-quran-phonetics" }, "dialects": [ "classical" ], "year": 2026, "notes": "Unlike standard Arabic ASR systems that output orthographic Arabic text, this model outputs a phonetic (sound-level) representation, making it suitable.", "base_model": [ "facebook/wav2vec2-base" ], "metrics": { "downloads": 8488, "likes": 9, "lastModified": "2026-01-21" } }, { "id": "arabic-handwritten-ocr-4bit-qwen2-5-vl-3b-v2", "name": "Arabic-handwritten-OCR-4bit-Qwen2.5-VL-3B-v2", "type": "ocr", "country": "INTL", "org": "sherif1313", "license": "apache-2.0", "modality": "vision", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/sherif1313/Arabic-handwritten-OCR-4bit-Qwen2.5-VL-3B-v2" }, "size": "3.9B", "year": 2025, "dialects": [ "msa" ], "notes": "4-bit Qwen2.5-VL 3B for Arabic handwritten OCR.", "base_model": [ "qwen/qwen2.5-vl-3b-instruct" ], "metrics": { "downloads": 8446, "likes": 12, "lastModified": "2025-12-29" } }, { "id": "tydiqa", "name": "TYDIQA", "type": "dataset", "country": "INTL", "org": "Google Research", "license": "apache-2.0", "modality": "text", "tasks": [ "qa" ], "links": { "github": "https://github.com/google-research-datasets/tydiqa", "hf": "https://huggingface.co/datasets/google-research-datasets/tydiqa", "paper": "https://aclanthology.org/2020.tacl-1.30.pdf" }, "dialects": [ "msa" ], "size": "25,893 sentences", "year": 2020, "tags": [ "multilingual" ], "notes": "Question answering dataset covering 11 typologically diverse languages with 200K question-answer pairs", "metrics": { "downloads": 7550, "likes": 38, "lastModified": "2024-08-08" } }, { "id": "arabic-triplet-matryoshka-v2", "name": "Arabic-Triplet-Matryoshka-V2", "type": "embedding", "country": "SA", "org": "Omartificial-Intelligence-Space", "license": "apache-2.0", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/Omartificial-Intelligence-Space/Arabic-Triplet-Matryoshka-V2" }, "notes": "Matryoshka representation", "base_model": [ "aubmindlab/bert-base-arabertv02" ], "metrics": { "downloads": 6958, "likes": 27, "lastModified": "2025-09-07" } }, { "id": "facebook-mms-tts-ara", "name": "facebook/mms-tts-ara", "type": "tts", "country": "INTL", "org": "Meta", "license": "cc-by-nc-4.0", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/facebook/mms-tts-ara" }, "on_device": true, "notes": "Facebook's Massively Multilingual Speech", "metrics": { "downloads": 6815, "likes": 21, "lastModified": "2023-09-01" } }, { "id": "wav2vec2-large-xlsr-53-arabic-quran-v-final", "name": "wav2vec2 large xlsr 53 arabic quran v final", "type": "asr", "country": "INTL", "org": "rabah2026", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr", "quran" ], "links": { "hf": "https://huggingface.co/rabah2026/wav2vec2-large-xlsr-53-arabic-quran-v_final" }, "dialects": [ "classical" ], "year": 2025, "notes": "Ce modèle est une version fine-tunée de jonatasgrosman/wav2vec2-large-xlsr-53-arabic sur le dataset rabah2026/Quran-Ayah-Corpus.", "metrics": { "downloads": 6448, "likes": 7, "lastModified": "2025-12-21" } }, { "id": "xquad", "name": "xquad", "type": "benchmark", "country": "INTL", "org": "HiTZ Center", "license": "cc-by-sa-4.0", "modality": "text", "tasks": [ "evaluation", "qa" ], "links": { "github": "https://github.com/deepmind/xquad", "hf": "https://huggingface.co/datasets/google/xquad", "paper": "https://aclanthology.org/2020.acl-main.421.pdf" }, "dialects": [ "msa" ], "size": "1,190 documents", "year": 2019, "tags": [ "multilingual" ], "notes": "Cross-lingual QA benchmark of 240 SQuAD paragraphs and 1190 question-answer pairs translated into ten languages including Arabic.", "metrics": { "downloads": 6365, "likes": 42, "lastModified": "2024-01-04" } }, { "id": "arabic-sts-matryoshka-v2", "name": "Arabic-STS-Matryoshka-V2", "type": "embedding", "country": "EG", "org": "Omar Elshehy", "license": "unknown", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/omarelshehy/Arabic-STS-Matryoshka-V2" }, "size": "135M", "year": 2024, "on_device": true, "dialects": [ "msa" ], "notes": "Arabic STS Matryoshka embedding model.", "base_model": [ "aubmindlab/bert-base-arabertv02" ], "metrics": { "downloads": 6240, "likes": 4, "lastModified": "2024-12-29" } }, { "id": "jais-2-8b-chat", "name": "Jais 2 8B Chat", "type": "llm", "country": "AE", "org": "Inception AI (G42)", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/inception42/Jais-2-8B-Chat", "paper": "https://arxiv.org/abs/2608.13580" }, "base_model": [ "from-scratch" ], "size": "8B", "year": 2025, "notes": "Smaller Jais 2 Arabic-English chat model with GGUF builds.", "metrics": { "downloads": 5822, "likes": 28, "lastModified": "2026-08-18" } }, { "id": "nemotron-3-5-quran-dual-v5", "name": "nemotron 3.5 quran dual v5", "type": "asr", "country": "INTL", "org": "tamm5y5m5", "license": "unknown", "modality": "speech", "tasks": [ "asr", "quran" ], "links": { "hf": "https://huggingface.co/tamm5y5m5/nemotron-3.5-quran-dual-v5" }, "dialects": [ "classical" ], "year": 2026, "notes": "Nemotron 3.5 streaming ASR (0.6B) fine-tuned for full-Quran recognition with tajweed marks; CER 2.76%.", "size": "0.6B", "on_device": true, "metrics": { "downloads": 5805, "likes": 2, "lastModified": "2026-09-02" } }, { "id": "universal-dependencies", "name": "Universal Dependencies", "type": "dataset", "country": "INTL", "org": "Universal Dependencies(UD)", "license": "unknown", "modality": "text", "tasks": [ "pos", "morphology" ], "links": { "website": "https://github.com/UniversalDependencies", "hf": "https://huggingface.co/datasets/universal-dependencies/universal_dependencies" }, "dialects": [ "msa" ], "size": "1,042,000 sentences", "year": 2020, "notes": "Universal Dependencies is a project that seeks to develop cross-linguistically consistent treebank annotation for many languages.", "metrics": { "downloads": 5742, "likes": 8, "lastModified": "2026-09-18" } }, { "id": "quranic-universal-ayahs", "name": "quranic universal ayahs", "type": "dataset", "country": "INTL", "org": "QUD-Technologies", "license": "cc-by-4.0", "modality": "speech", "tasks": [ "asr", "alignment" ], "links": { "hf": "https://huggingface.co/datasets/QUD-Technologies/quranic-universal-ayahs" }, "dialects": [ "classical" ], "size": "100K–1M rows", "year": 2026, "notes": "Qur'anic ayah recitation audio with forced-alignment word and letter timestamps for ASR and segmentation.", "metrics": { "downloads": 5584, "likes": 7, "lastModified": "2026-10-04" } }, { "id": "lahgtna-v2", "name": "Dialectal Arabic Lahgtna v2", "type": "dataset", "country": "EG", "org": "oddadmix", "license": "unknown", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/datasets/oddadmix/dialectal-arabic-lahgtna-v2" }, "size": "3000h", "year": 2026, "dialects": [ "mixed" ], "notes": "Multi-dialect Arabic speech across 13 dialects, 3,000+ hours.", "metrics": { "downloads": 5530, "likes": 29, "lastModified": "2026-08-04" } }, { "id": "audar-asr-v1-turbo", "name": "Audar ASR V1 Turbo", "type": "asr", "country": "INTL", "org": "Audar AI", "license": "other", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/audarai/Audar-ASR-V1-Turbo" }, "dialects": [ "gulf" ], "year": 2026, "notes": "Audar's Arabic-first speech-recognition model — leaderboard-grade, dialect-aware.", "metrics": { "downloads": 5469, "likes": 17, "lastModified": "2026-08-20" } }, { "id": "arabic-pp-ocrv5-mobile-rec", "name": "arabic_PP-OCRv5_mobile_rec", "type": "ocr", "country": "INTL", "org": "PaddlePaddle", "license": "apache-2.0", "modality": "vision", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/PaddlePaddle/arabic_PP-OCRv5_mobile_rec" }, "year": 2025, "on_device": true, "dialects": [ "msa" ], "notes": "PP-OCRv5 mobile recognition model for Arabic script.", "metrics": { "downloads": 5105, "likes": 7, "lastModified": "2025-10-16" } }, { "id": "quranic-translation-audio-data", "name": "Quranic Translation Audio Data", "type": "dataset", "country": "INTL", "org": "zaibihassan", "license": "apache-2.0", "modality": "speech", "tasks": [ "tts", "quran" ], "links": { "hf": "https://huggingface.co/datasets/zaibihassan/Quranic-Translation-Audio-Data" }, "dialects": [ "classical" ], "size": "1K–10K rows", "year": 2026, "notes": "Complete Quran translation audio and commentary recitations across 51 translation directories in 14 languages.", "metrics": { "downloads": 5082, "likes": 1, "lastModified": "2026-10-05" } }, { "id": "arabic-retrieval-v1-0", "name": "Arabic-Retrieval-v1.0", "type": "embedding", "country": "EG", "org": "Omar Elshehy", "license": "apache-2.0", "modality": "text", "tasks": [ "embedding", "retrieval" ], "links": { "hf": "https://huggingface.co/omarelshehy/Arabic-Retrieval-v1.0" }, "size": "135M", "year": 2024, "on_device": true, "dialects": [ "msa" ], "notes": "Arabic retrieval embedding model.", "base_model": [ "aubmindlab/bert-base-arabertv02" ], "metrics": { "downloads": 5061, "likes": 6, "lastModified": "2024-12-19" } }, { "id": "prophet-mosque-library-compressed", "name": "prophet mosque library compressed", "type": "dataset", "country": "INTL", "org": "ieasybooks", "license": "mit", "modality": "text", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/datasets/ieasybooks-org/prophet-mosque-library-compressed" }, "size": "10K–100K rows", "year": 2025, "notes": "Prophet’s Mosque Library is one of the primary resources for Islamic books.", "metrics": { "downloads": 4957, "likes": 0, "lastModified": "2025-05-08" } }, { "id": "quranic-word-by-word-audio-data", "name": "Quranic Word By Word Audio Data", "type": "dataset", "country": "INTL", "org": "zaibihassan", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr", "tts" ], "links": { "hf": "https://huggingface.co/datasets/zaibihassan/Quranic-Word-By-Word-Audio-Data" }, "dialects": [ "classical" ], "size": "100K–1M rows", "year": 2026, "notes": "Two word-by-word Quran recitation sets, Muallim (teacher style) and Mujawwad (tajweed style), for ASR and TTS.", "metrics": { "downloads": 4876, "likes": 4, "lastModified": "2026-05-23" } }, { "id": "waqfeya-library-compressed", "name": "waqfeya library compressed", "type": "dataset", "country": "INTL", "org": "ieasybooks", "license": "mit", "modality": "text", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/datasets/ieasybooks-org/waqfeya-library-compressed" }, "size": "10K–100K rows", "year": 2025, "notes": "Waqfeya is one of the primary online resources for Islamic books, similar to Shamela.", "metrics": { "downloads": 4875, "likes": 6, "lastModified": "2025-04-25" } }, { "id": "arwiki", "name": "arwiki", "type": "dataset", "country": "INTL", "org": "CALM", "license": "['unknown']", "modality": "text", "tasks": [ "nlp" ], "links": { "hf": "https://huggingface.co/datasets/CALM/arwiki" }, "dialects": [ "msa" ], "size": "10M–100M rows", "year": 2022, "notes": "This dataset is extracted using wikiextractor tool, from Wikipedia Arabic pages.", "metrics": { "downloads": 4509, "likes": 5, "lastModified": "2022-08-01" } }, { "id": "bert-base-arabertv02-twitter", "name": "bert-base-arabertv02-twitter", "type": "llm", "country": "LB", "org": "AUB MIND Lab", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/aubmindlab/bert-base-arabertv02-twitter" }, "size": "135M", "year": 2022, "on_device": true, "dialects": [ "mixed" ], "notes": "AraBERT v02 base further pretrained on tweets including dialects.", "metrics": { "downloads": 4198, "likes": 8, "lastModified": "2023-03-23" } }, { "id": "miracl", "name": "MIRACL", "type": "dataset", "country": "INTL", "org": "unknown", "license": "apache-2.0", "modality": "text", "tasks": [ "retrieval" ], "links": { "hf": "https://huggingface.co/datasets/miracl/miracl-corpus" }, "dialects": [ "msa" ], "size": "2,061,414 documents", "year": 2022, "tags": [ "multilingual" ], "notes": "MIRACL multilingual retrieval corpus spanning 18 languages, with Arabic as one documented subset.", "metrics": { "downloads": 4116, "likes": 54, "lastModified": "2023-01-05" } }, { "id": "everyayah-tarteel", "name": "Tarteel EveryAyah", "type": "dataset", "country": "INTL", "org": "Tarteel AI", "license": "unknown", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/datasets/tarteel-ai/everyayah" }, "year": 2022, "dialects": [ "classical" ], "notes": "Official Tarteel EveryAyah dataset of Quran recitations by multiple reciters.", "metrics": { "downloads": 4016, "likes": 42, "lastModified": "2026-09-17" } }, { "id": "ontonotes-5-0", "name": "OntoNotes 5.0", "type": "dataset", "country": "INTL", "org": "Raytheon BBN Technologies", "license": "proprietary", "modality": "text", "tasks": [ "ner", "coreference" ], "links": { "website": "https://catalog.ldc.upenn.edu/LDC2013T19", "hf": "https://huggingface.co/datasets/ontonotes/conll2012_ontonotesv5", "paper": "https://catalog.ldc.upenn.edu/docs/LDC2013T19/OntoNotes-Release-5.0.pdf" }, "dialects": [ "msa" ], "size": "300,000 tokens", "year": 2012, "tags": [ "multilingual", "variants:2" ], "notes": "Arabic portion of OntoNotes 5.0, with 300K words of newswire annotated by LDC.", "metrics": { "downloads": 3958, "likes": 46, "lastModified": "2024-01-18" } }, { "id": "arbertv2", "name": "ARBERTv2", "type": "llm", "country": "INTL", "org": "UBC-NLP", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/UBC-NLP/ARBERTv2" }, "size": "164M", "year": 2023, "on_device": true, "dialects": [ "msa" ], "notes": "ARBERT v2, MSA BERT from UBC.", "metrics": { "downloads": 3906, "likes": 9, "lastModified": "2024-04-24" } }, { "id": "arat5", "name": "AraT5", "type": "llm", "country": "INTL", "org": "UBC-NLP", "license": "unknown", "modality": "text", "tasks": [ "encoder" ], "links": { "hf": "https://huggingface.co/UBC-NLP/AraT5-base" }, "notes": "T5 for Arabic summarization, translation, paraphrasing", "metrics": { "downloads": 3854, "likes": 22, "lastModified": "2024-05-16" } }, { "id": "qurantts", "name": "QuranTTS", "type": "dataset", "country": "INTL", "org": "Quran Lab", "license": "other", "modality": "speech", "tasks": [ "tts", "restoration" ], "links": { "hf": "https://huggingface.co/datasets/Quran-Lab/QuranTTS" }, "dialects": [ "classical" ], "size": "10K–100K rows", "year": 2026, "notes": "QuranTTS v4: ear-verified Quranic recitation corpus at 24-bit/48 kHz for TTS and speech restoration; non-commercial.", "metrics": { "downloads": 3830, "likes": 4, "lastModified": "2026-09-21" } }, { "id": "arabic-stem-lexicon", "name": "arabic stem lexicon", "type": "dataset", "country": "INTL", "org": "TigreGotico", "license": "cc-by-4.0", "modality": "text", "tasks": [ "tts", "diacritization" ], "links": { "hf": "https://huggingface.co/datasets/TigreGotico/arabic-stem-lexicon" }, "size": "100K–1M rows", "year": 2026, "notes": "An undiacritized Arabic surface form → its most frequent diacritized stem.", "metrics": { "downloads": 3773, "likes": 0, "lastModified": "2026-07-14" } }, { "id": "tunisianmmlu", "name": "TunisianMMLU", "type": "benchmark", "country": "INTL", "org": "LINAGORA", "license": "cc-by-nc-sa-4.0", "modality": "text", "tasks": [ "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/linagora/TunisianMMLU" }, "size": "10K-100K rows", "year": 2025, "dialects": [ "magh" ], "notes": "MMLU translated into Tunisian Derja, LINAGORA (France); evaluated with lighteval.", "metrics": { "downloads": 3724, "likes": 0, "lastModified": "2026-09-21" } }, { "id": "whisper-large-v3-ar", "name": "whisper large v3 ar", "type": "asr", "country": "INTL", "org": "Dr-AliGomaa", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr", "quran", "hadith" ], "links": { "hf": "https://huggingface.co/Dr-AliGomaa/whisper-large-v3-ar" }, "dialects": [ "egy", "classical", "msa" ], "year": 2026, "notes": "openai/whisper-large-v3 fine-tuned for Quran and Hadith transcription without giving up general Modern Standard Arabic.", "base_model": [ "openai/whisper-large-v3" ], "metrics": { "downloads": 3719, "likes": 0, "lastModified": "2026-08-18" } }, { "id": "bert-base-arabic-camelbert-msa", "name": "bert base arabic camelbert msa", "type": "llm", "country": "AE", "org": "CAMeL Lab, NYUAD", "license": "apache-2.0", "modality": "text", "tasks": [ "encoder", "pretraining" ], "links": { "hf": "https://huggingface.co/CAMeL-Lab/bert-base-arabic-camelbert-msa" }, "dialects": [ "classical", "msa" ], "year": 2021, "notes": "We release pre-trained language models for Modern Standard Arabic (MSA), dialectal Arabic (DA), and classical Arabic.", "metrics": { "downloads": 3648, "likes": 10, "lastModified": "2021-09-14" } }, { "id": "arabicmmlu", "name": "ArabicMMLU", "type": "benchmark", "country": "AE", "org": "MBZUAI", "license": "cc-by-nc-4.0", "modality": "text", "tasks": [ "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/MBZUAI/ArabicMMLU" }, "notes": "Multi-task language understanding from school exams", "metrics": { "downloads": 3623, "likes": 39, "lastModified": "2024-09-17" } }, { "id": "common-voice-arabic", "name": "Common Voice Arabic", "type": "dataset", "country": "INTL", "org": "Mozilla Foundation", "license": "cc0-1.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/datasets/mozilla-foundation/common_voice_17_0", "website": "https://commonvoice.mozilla.org/ar/datasets" }, "year": 2017, "dialects": [ "mixed" ], "notes": "Crowdsourced Arabic read speech in Common Voice, standard ASR training and evaluation set.", "metrics": { "downloads": 3532, "likes": 52, "lastModified": "2025-10-24" } }, { "id": "muaalem-annotated-v3", "name": "muaalem annotated v3", "type": "dataset", "country": "INTL", "org": "obadx", "license": "mit", "modality": "speech", "tasks": [ "quran" ], "links": { "hf": "https://huggingface.co/datasets/obadx/muaalem-annotated-v3" }, "dialects": [ "classical" ], "size": "100K–1M rows", "year": 2025, "notes": "Quran recitations from expert reciters with phonetic transcripts encoding tajweed rules, for recitation-error detection models.", "metrics": { "downloads": 3524, "likes": 7, "lastModified": "2025-09-04" } }, { "id": "ara-reranker-v1", "name": "ARA-Reranker-V1", "type": "embedding", "country": "SA", "org": "Omartificial-Intelligence-Space", "license": "apache-2.0", "modality": "text", "tasks": [ "reranking" ], "links": { "hf": "https://huggingface.co/Omartificial-Intelligence-Space/ARA-Reranker-V1" }, "size": "568M", "year": 2024, "on_device": true, "dialects": [ "msa" ], "notes": "Arabic cross-encoder reranker.", "metrics": { "downloads": 3391, "likes": 4, "lastModified": "2025-04-03" } }, { "id": "unnamed-split-ar", "name": "unnamed split ar", "type": "dataset", "country": "EG", "org": "oddadmix", "license": "cc-by-3.0", "modality": "speech", "tasks": [ "asr", "speech" ], "links": { "hf": "https://huggingface.co/datasets/oddadmix/unnamed-split-ar" }, "year": 2026, "notes": "The data/ar/ partition of espnet/yodas3, cut from long recordings (median 5 minutes) into clips of 5-45 seconds, each with its own transcript.", "metrics": { "downloads": 3247, "likes": 0, "lastModified": "2026-10-05" } }, { "id": "bert-large-arabertv02", "name": "bert-large-arabertv02", "type": "llm", "country": "LB", "org": "AUB MIND Lab", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/aubmindlab/bert-large-arabertv02" }, "size": "371M", "year": 2022, "on_device": true, "dialects": [ "msa" ], "notes": "AraBERT v02 large encoder.", "metrics": { "downloads": 3246, "likes": 10, "lastModified": "2023-08-03" } }, { "id": "menaspeechbank", "name": "MENASpeechBank", "type": "dataset", "country": "QA", "org": "QCRI", "license": "cc-by-nc-sa-4.0", "modality": "speech", "tasks": [ "asr", "tts", "chat" ], "links": { "hf": "https://huggingface.co/datasets/QCRI/MenaSpeechBank" }, "size": "26.5h", "year": 2026, "dialects": [ "mixed" ], "notes": "MENA speech resource with reference voice bank and synthetic multi-turn persona dialogues.", "metrics": { "downloads": 3184, "likes": 4, "lastModified": "2026-08-28" } }, { "id": "qwen3-5-9b-saudi-dialect", "name": "Qwen3.5 9B saudi dialect", "type": "llm", "country": "INTL", "org": "AyoubChLin", "license": "unknown", "modality": "text", "tasks": [ "chat", "instruction-tuning", "vision" ], "links": { "hf": "https://huggingface.co/AyoubChLin/Qwen3.5-9B-saudi-dialect" }, "dialects": [ "gulf" ], "size": "9B", "on_device": false, "year": 2026, "tags": [ "variants:2" ], "notes": "This repository contains merged full weights for Saudi-dialect chat generation, not just LoRA adapters. (also: 2 variants)", "base_model": [ "unsloth/qwen3.5-9b" ], "metrics": { "downloads": 3177, "likes": 0, "lastModified": "2026-03-22" } }, { "id": "egyptian-arabic-wav2vec2-xlsr-53", "name": "egyptian-arabic-wav2vec2-xlsr-53", "type": "asr", "country": "EG", "org": "Ibrahim Amin", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/IbrahimAmin/egyptian-arabic-wav2vec2-xlsr-53" }, "size": "315M", "year": 2025, "on_device": true, "dialects": [ "egy" ], "notes": "wav2vec2 XLSR-53 fine-tuned on Egyptian Arabic.", "base_model": [ "omarxadel/wav2vec2-large-xlsr-53-arabic-egyptian" ], "metrics": { "downloads": 3130, "likes": 6, "lastModified": "2025-05-11" } }, { "id": "opus-mt-tc-big-en-ar", "name": "opus mt tc big en ar", "type": "llm", "country": "INTL", "org": "Helsinki-NLP", "license": "cc-by-4.0", "modality": "text", "tasks": [ "translation" ], "links": { "hf": "https://huggingface.co/Helsinki-NLP/opus-mt-tc-big-en-ar" }, "year": 2023, "notes": "Neural machine translation model for translating from English (en) to Arabic (ar).", "metrics": { "downloads": 3086, "likes": 29, "lastModified": "2023-10-10" } }, { "id": "masri-100h", "name": "Masri 100h", "type": "dataset", "country": "EG", "org": "Ehab Negm", "license": "cc-by-nc-4.0", "modality": "speech", "tasks": [ "asr", "tts" ], "links": { "hf": "https://huggingface.co/datasets/ehabnegm/100-hour-Egyptian-dataset-single-speaker" }, "size": "100h", "year": 2026, "dialects": [ "egy" ], "notes": "100h Egyptian Arabic single-narrator speech corpus, 15.6k released clips.", "metrics": { "downloads": 3078, "likes": 11, "lastModified": "2026-08-07" } }, { "id": "yemeni-arabic-assistant", "name": "Gemma-4 E2B Yemeni Arabic Assistant", "type": "llm", "country": "YE", "org": "Yemeni AI Lab", "license": "cc-by-nc-sa-4.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/yemeni-ai-lab/gemma-4-e2b-yemeni-arabic-assistant-GGUF" }, "dialects": [ "yemeni" ], "notes": "Small Gemma-based chat assistant for Yemeni dialect, GGUF (Yemen).", "base_model": [ "unsloth/gemma-4-e2b-it", "google/gemma-4-e2b-it" ], "metrics": { "downloads": 3058, "likes": 0, "lastModified": "2026-05-04" } }, { "id": "aragpt2-base", "name": "aragpt2-base", "type": "llm", "country": "LB", "org": "AUB MIND Lab", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/aubmindlab/aragpt2-base" }, "base_model": [ "from-scratch" ], "size": "148M", "year": 2022, "on_device": true, "dialects": [ "msa" ], "notes": "AraGPT2 base, GPT-2 style Arabic generator from AUB.", "metrics": { "downloads": 2927, "likes": 34, "lastModified": "2023-10-30" } }, { "id": "masribert-v4", "name": "MASRIBERT v4", "type": "llm", "country": "EG", "org": "T0KII", "license": "unknown", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/T0KII/MASRIBERTV4" }, "size": "240M", "dialects": [ "egy" ], "notes": "BERT-style encoder for Egyptian Arabic (Masri) text; community model.", "base_model": [ "ubc-nlp/marbertv2" ], "metrics": { "downloads": 2914, "likes": 0, "lastModified": "2026-08-17" } }, { "id": "belebele-fleurs", "name": "Belebele-Fleurs", "type": "benchmark", "country": "INTL", "org": "University of Würzburg", "license": "cc-by-sa-4.0", "modality": "speech", "tasks": [ "evaluation", "qa" ], "links": { "hf": "https://huggingface.co/datasets/WueNLP/belebele-fleurs", "paper": "https://arxiv.org/pdf/2501.06117" }, "dialects": [ "msa" ], "size": "387 sentences", "year": 2025, "tags": [ "multilingual" ], "notes": "Belebele-Fleurs extends BeleBele into a spoken SLU benchmark with multilingual multiple-choice listening comprehension QA from speech.", "metrics": { "downloads": 2888, "likes": 9, "lastModified": "2024-12-12" } }, { "id": "qari-ocr-v0-3-vl-2b-instruct", "name": "Qari-OCR-v0.3-VL-2B-Instruct", "type": "ocr", "country": "SA", "org": "NAMAA-Space", "license": "apache-2.0", "modality": "vision", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/NAMAA-Space/Qari-OCR-v0.3-VL-2B-Instruct" }, "size": "2.2B", "year": 2025, "on_device": true, "dialects": [ "msa" ], "notes": "Qari OCR v0.3, 2B Qwen2-VL based Arabic OCR with diacritics support.", "metrics": { "downloads": 2831, "likes": 22, "lastModified": "2025-06-10" } }, { "id": "darija-mmlu", "name": "DarijaMMLU", "type": "benchmark", "country": "AE", "org": "MBZUAI-Paris Lab", "license": "mit", "modality": "text", "tasks": [ "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/MBZUAI-Paris/DarijaMMLU", "paper": "https://arxiv.org/abs/2409.17912" }, "size": "10K-100K rows", "year": 2024, "dialects": [ "magh" ], "notes": "MMLU translated into Moroccan Darija, 22k+ multiple-choice questions.", "metrics": { "downloads": 2807, "likes": 8, "lastModified": "2024-09-27" } }, { "id": "lahgtna-libyan-tts-dataset", "name": "Lahgtna Libyan TTS dataset", "type": "dataset", "country": "INTL", "org": "oddadmix", "license": "unknown", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/datasets/oddadmix/lahgtna-libyan-tts" }, "dialects": [ "magh" ], "year": 2026, "notes": "Libyan Arabic text-to-speech dataset from the Lahgtna multi-dialect TTS project.", "metrics": { "downloads": 2702, "likes": 0, "lastModified": "2026-08-31" } }, { "id": "arabic-aya", "name": "Arabic Aya (2A)", "type": "dataset", "country": "INTL", "org": "2A2I", "license": "apache-2.0", "modality": "text", "tasks": [ "instruction-tuning" ], "links": { "hf": "https://huggingface.co/datasets/2A2I/Arabic_Aya" }, "size": "1M-10M rows", "year": 2024, "dialects": [ "msa" ], "notes": "Arabic subset and translations of the Aya collection for instruction tuning.", "metrics": { "downloads": 2692, "likes": 16, "lastModified": "2024-03-15" } }, { "id": "araelectra-base-artydiqa", "name": "araelectra-base-artydiqa", "type": "llm", "country": "LB", "org": "Wissam Antoun (AUB)", "license": "unknown", "modality": "text", "tasks": [ "question-answering" ], "links": { "hf": "https://huggingface.co/wissamantoun/araelectra-base-artydiqa" }, "size": "135M", "year": 2022, "on_device": true, "dialects": [ "msa" ], "notes": "AraELECTRA base fine-tuned on Arabic TyDiQA.", "metrics": { "downloads": 2686, "likes": 12, "lastModified": "2023-03-20" } }, { "id": "mixed-arabic-datasets-mad", "name": "Mixed Arabic Datasets (MAD)", "type": "dataset", "country": "INTL", "org": "M-A-D", "license": "unknown", "modality": "text", "tasks": [ "nlp" ], "links": { "hf": "https://huggingface.co/datasets/M-A-D/Mixed-Arabic-Datasets-Repo" }, "notes": "Community-driven diverse Arabic texts", "metrics": { "downloads": 2664, "likes": 38, "lastModified": "2023-10-16" } }, { "id": "dialectal-arabic-voices", "name": "Dialectal Arabic Voices", "type": "dataset", "country": "INTL", "org": "moaead", "license": "unknown", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/datasets/moaead/dialectal-arabic-voices" }, "year": 2026, "dialects": [ "lev" ], "notes": "60,775 untranscribed Arabic recordings, currently labelled Palestinian Arabic.", "metrics": { "downloads": 2630, "likes": 0, "lastModified": "2026-09-26" } }, { "id": "araelectra", "name": "AraELECTRA", "type": "llm", "country": "LB", "org": "AUB MIND Lab", "license": "unknown", "modality": "text", "tasks": [ "encoder" ], "links": { "hf": "https://huggingface.co/aubmindlab/araelectra-base-discriminator" }, "notes": "ELECTRA for Arabic", "metrics": { "downloads": 2540, "likes": 5, "lastModified": "2024-10-29" } }, { "id": "arabic-sts", "name": "arabic sts", "type": "dataset", "country": "INTL", "org": "MohamedRashad", "license": "unknown", "modality": "text", "tasks": [ "nlp" ], "links": { "hf": "https://huggingface.co/datasets/MohamedRashad/arabic-sts" }, "size": "10K–100K rows", "year": 2024, "notes": "In the time of writing this Dataset Card, 31,112 civilians has been killed in Gaza (2/3 of them are women, elderly and children).", "metrics": { "downloads": 2480, "likes": 6, "lastModified": "2024-03-17" } }, { "id": "ar-stablelm-2-chat", "name": "Ar-stablelm-2-chat", "type": "llm", "country": "SA", "org": "Stability AI", "license": "other", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/stabilityai/stablelm-2-1_6b-chat" }, "size": "1.6B", "notes": "Small Arabic chat model", "on_device": true, "metrics": { "downloads": 2469, "likes": 34, "lastModified": "2024-06-03" } }, { "id": "101-billion-arabic-words", "name": "101 Billion Arabic Words", "type": "dataset", "country": "INTL", "org": "ClusterlabAi", "license": "apache-2.0", "modality": "text", "tasks": [ "nlp" ], "links": { "hf": "https://huggingface.co/datasets/ClusterlabAi/101_billion_arabic_words_dataset" }, "notes": "Massive Arabic web corpus", "metrics": { "downloads": 2416, "likes": 73, "lastModified": "2024-06-16" } }, { "id": "ara-prompt-guard-v0", "name": "Ara-Prompt-Guard_V0", "type": "llm", "country": "SA", "org": "NAMAA-Space", "license": "apache-2.0", "modality": "text", "tasks": [ "classification" ], "links": { "hf": "https://huggingface.co/NAMAA-Space/Ara-Prompt-Guard_V0" }, "size": "279M", "year": 2025, "on_device": true, "dialects": [ "msa" ], "notes": "Arabic prompt-injection and jailbreak classifier from NAMAA.", "base_model": [ "meta-llama/llama-prompt-guard-2-86m" ], "metrics": { "downloads": 2405, "likes": 11, "lastModified": "2026-03-08" } }, { "id": "ocr-arabic-books", "name": "ocr arabic books", "type": "dataset", "country": "INTL", "org": "freococo", "license": "cc-by-nc-nd-4.0", "modality": "multimodal", "tasks": [ "ocr", "vision" ], "links": { "hf": "https://huggingface.co/datasets/freococo/ocr_arabic_books" }, "size": "100K–1M rows", "year": 2026, "notes": "This repository is a structurally aligned Arabic Optical Character Recognition (OCR) dataset.", "metrics": { "downloads": 2380, "likes": 5, "lastModified": "2026-07-09" } }, { "id": "2m-belebele", "name": "2M-BELEBELE", "type": "dataset", "country": "INTL", "org": "Meta", "license": "cc-by-sa-4.0", "modality": "speech", "tasks": [ "speech-comprehension" ], "links": { "github": "https://github.com/facebookresearch/belebele", "hf": "https://huggingface.co/datasets/facebook/2M-Belebele", "paper": "https://doi.org/10.18653/v1/2025.findings-acl.569" }, "dialects": [ "msa" ], "size": "2,000 sentences", "year": 2024, "tags": [ "multilingual" ], "notes": "Multilingual speech and ASL comprehension dataset extending BELEBELE to 75 languages.", "metrics": { "downloads": 2355, "likes": 13, "lastModified": "2024-12-17" } }, { "id": "bert-base-arabic-camelbert-ca", "name": "bert-base-arabic-camelbert-ca", "type": "llm", "country": "AE", "org": "CAMeL Lab, NYUAD", "license": "apache-2.0", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/CAMeL-Lab/bert-base-arabic-camelbert-ca" }, "year": 2022, "dialects": [ "classical" ], "notes": "CAMeLBERT-CA pretrained on Classical Arabic.", "metrics": { "downloads": 2342, "likes": 13, "lastModified": "2021-09-14" } }, { "id": "mlqa", "name": "MLQA", "type": "dataset", "country": "INTL", "org": "Facebook AI Research", "license": "cc-by-sa-3.0", "modality": "text", "tasks": [ "qa" ], "links": { "github": "https://github.com/facebookresearch/mlqa", "hf": "https://huggingface.co/datasets/facebook/mlqa", "paper": "https://arxiv.org/pdf/1910.07475.pdf" }, "dialects": [ "msa" ], "size": "5,852 documents", "year": 2020, "tags": [ "multilingual" ], "notes": "5K extractive QA instances (12K in English) in SQuAD format in seven languages - English, Arabic, German, Spanish, Hindi, Vietnamese and Simplified Chinese.", "metrics": { "downloads": 2305, "likes": 44, "lastModified": "2024-01-18" } }, { "id": "bert-base-arabic-camelbert-da", "name": "bert-base-arabic-camelbert-da", "type": "llm", "country": "AE", "org": "CAMeL Lab, NYUAD", "license": "apache-2.0", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/CAMeL-Lab/bert-base-arabic-camelbert-da" }, "year": 2022, "dialects": [ "mixed" ], "notes": "CAMeLBERT-DA pretrained on dialectal Arabic.", "metrics": { "downloads": 2300, "likes": 28, "lastModified": "2021-09-14" } }, { "id": "masc-arabic", "name": "MASC Arabic", "type": "dataset", "country": "INTL", "org": "MohamedRashad", "license": "cc-by-4.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/datasets/MohamedRashad/MASC-Arabic" }, "size": "1000h", "year": 2023, "dialects": [ "mixed" ], "notes": "Massive Arabic Speech Corpus, about 1,000h of YouTube speech across dialects, on HF.", "metrics": { "downloads": 2258, "likes": 8, "lastModified": "2026-04-06" } }, { "id": "mizan-rerank-v2", "name": "Mizan Rerank V2", "type": "embedding", "country": "INTL", "org": "ALJIACHI", "license": "apache-2.0", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/ALJIACHI/Mizan-Rerank-V2" }, "year": 2026, "tags": [ "variants:2" ], "notes": "A cross-encoder model for reranking Arabic long texts, fine-tuned from Alibaba-NLP/gte-multilingual-reranker-base. (also: 2 variants)", "base_model": [ "alibaba-nlp/gte-multilingual-reranker-base" ], "metrics": { "downloads": 2128, "likes": 0, "lastModified": "2026-09-29" } }, { "id": "cohere-command-a", "name": "Cohere Command-A", "type": "llm", "country": "INTL", "org": "Cohere", "license": "cc-by-nc-4.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/CohereForAI/c4ai-command-a-03-2025" }, "size": "111B", "notes": "Optimized for RAG", "metrics": { "downloads": 2124, "likes": 395, "lastModified": "2025-10-30" } }, { "id": "arat5v2-base-1024", "name": "AraT5v2-base-1024", "type": "llm", "country": "INTL", "org": "UBC-NLP", "license": "unknown", "modality": "text", "tasks": [ "translation", "pretraining" ], "links": { "hf": "https://huggingface.co/UBC-NLP/AraT5v2-base-1024" }, "year": 2023, "dialects": [ "msa" ], "notes": "AraT5 v2 base, Arabic text-to-text transformer with 1024 context.", "metrics": { "downloads": 2041, "likes": 32, "lastModified": "2024-05-16" } }, { "id": "aramix-hq", "name": "AraMix-HQ", "type": "dataset", "country": "INTL", "org": "AdaMLLab", "license": "other", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/datasets/AdaMLLab/AraMix-HQ" }, "size": "10M–100M rows", "year": 2026, "notes": "AraMix-HQ: 53.7M Arabic text rows from AraMix filtered with a model quality score (mmbert_score).", "metrics": { "downloads": 2038, "likes": 2, "lastModified": "2026-01-30" } }, { "id": "wav2vec2-large-xlsr-moroccan-darija", "name": "wav2vec2-large-xlsr-moroccan-darija", "type": "asr", "country": "MA", "org": "Boumehdi", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/boumehdi/wav2vec2-large-xlsr-moroccan-darija" }, "year": 2023, "dialects": [ "magh" ], "notes": "wav2vec2 XLSR fine-tuned on Moroccan Darija.", "base_model": [ "facebook/wav2vec2-large-xlsr-53" ], "metrics": { "downloads": 2031, "likes": 32, "lastModified": "2024-04-12" } }, { "id": "sada22", "name": "SADA22", "type": "dataset", "country": "INTL", "org": "MohamedRashad", "license": "cc-by-nc-sa-4.0", "modality": "speech", "tasks": [ "asr", "tts", "speech" ], "links": { "hf": "https://huggingface.co/datasets/MohamedRashad/SADA22" }, "dialects": [ "gulf", "mixed" ], "size": "100K–1M rows", "year": 2025, "notes": "The SADA dataset (Saudi Audio Dataset for Arabic) is a large-scale Arabic speech corpus designed to support the development.", "metrics": { "downloads": 2029, "likes": 30, "lastModified": "2025-05-03" } }, { "id": "systematic-review-and-meta-analysis-of-2024-2025-studies-on", "name": "Systematic Review and Meta-Analysis of 2024–2025 Studies on AI Arabic Translation, Linguistics and Pedagogy", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "translation" ], "links": { "paper": "https://doi.org/10.32996/jcsts.2026.5.1.2" }, "year": 2026, "venue": "Frontiers in Computer Science and Artificial Intelligence", "citations": 30, "notes": "This study aims to conduct a systematic review (SR) and meta-analysis (MA) of twenty articles by the author published between 2024–2025 on the use of AI models", "metrics": null }, { "id": "a-holistic-assessment-of-the-carbon-footprint-of-noor-a-very", "name": "A Holistic Assessment of the Carbon Footprint of Noor, a Very Large Arabic Language Model", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2610.00223" }, "year": 2026, "venue": "BIGSCIENCE", "citations": 26, "notes": "As ever larger language models grow more ubiquitous, it is crucial to consider their environmental impact.", "metrics": null }, { "id": "a-comparative-study-of-pretrained-transformer-models-for-quranic-asr-s", "name": "A Comparative Study of Pretrained Transformer Models for Quranic ASR: Speech Representations, Label Formats, and Dataset Composition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "diacritization", "pretraining", "speech", "quran" ], "links": { "paper": "https://arxiv.org/abs/2606.19747" }, "dialects": [ "classical" ], "year": 2026, "venue": "arXiv 2026", "notes": "This paper presents a systematic empirical study of domain-specific fine-tuning of pretrained Transformer-based models for Quranic ASR.", "metrics": null }, { "id": "a-corpus-aligned-uthmani-to-standard-quranic-word-mapping-and-a-determ", "name": "A Corpus-Aligned Uthmani-to-Standard Quranic Word Mapping and a Deterministic Recitation Validator", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "evaluation", "quran" ], "links": { "paper": "https://arxiv.org/abs/2609.14967" }, "dialects": [ "classical" ], "year": 2026, "venue": "arXiv 2026", "notes": "On top of the normalized text, we build a deterministic, LLM-free Quranic recitation validator using a four-layer verse-matching search.", "metrics": null }, { "id": "a-human-in-the-loop-label-error-detection-framework-applied-to-arabic", "name": "A Human-in-the-Loop Label Error Detection Framework Applied to Arabic-Script HTR Datasets", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "asr", "ocr", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2601.16713" }, "year": 2026, "venue": "arXiv 2026", "notes": "To help closing this gap, we propose a two-stage framework (CER-HV) for detecting label errors.", "metrics": null }, { "id": "abjad-kids-an-arabic-speech-classification-dataset-for-primary-educati", "name": "Abjad-Kids: An Arabic Speech Classification Dataset for Primary Education", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2603.20255" }, "year": 2026, "venue": "arXiv 2026", "notes": "However, children speech research remains limited due to the lack of publicly available datasets.", "metrics": null }, { "id": "adab-arabic-dataset-for-automated-politeness-benchmarking-a-large-scal", "name": "ADAB: Arabic Dataset for Automated Politeness Benchmarking -- A Large-Scale Resource for Computational Sociopragmatics", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2602.13870" }, "dialects": [ "egy", "gulf", "lev", "magh", "msa", "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "We introduce ADAB (Arabic Politeness Dataset), a new annotated Arabic dataset collected from four online platforms, including social media, e-commerce.", "metrics": null }, { "id": "aha-memes-a-fine-grained-multimodal-benchmark-for-understanding-hate-i", "name": "AHA-Memes: A Fine-Grained Multimodal Benchmark for Understanding Hate in Arabic Memes", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2607.27393" }, "year": 2026, "venue": "arXiv 2026", "notes": "We introduce AHA-Memes (Arabic HAteful Memes), which is, to our knowledge, the first large-scale Arabic hateful meme benchmark with fine-grained.", "metrics": null }, { "id": "alexandria-a-multi-domain-dialectal-arabic-machine-translation-dataset", "name": "Alexandria: A Multi-Domain Dialectal Arabic Machine Translation Dataset for Culturally Inclusive and Linguistically Diverse LLMs", "type": "paper", "country": "INTL", "org": "UBC-NLP", "license": "unknown", "modality": "text", "tasks": [ "chat", "translation", "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2601.13099", "github": "https://github.com/UBC-NLP/Alexandria" }, "dialects": [ "msa", "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "We introduce Alexandria, a large-scale, community-driven, human-translated dataset designed to bridge this gap.", "metrics": null }, { "id": "alexandriax-2026-the-first-shared-task-on-dialectal-arabic-machine-tra", "name": "AlexandriaX 2026: The First Shared Task on Dialectal Arabic Machine Translation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "chat", "embedding", "translation", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2609.22796", "website": "https://alexandriax.dlnlp.ai" }, "dialects": [ "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "We present the AlexandriaX 2026 Shared Task on Dialectal Arabic MT, which addresses these challenges through three complementary subtasks.", "metrics": null }, { "id": "almieyar-oryx-bloombench-a-bilingual-multimodal-benchmark-for-cognitiv", "name": "Almieyar-Oryx-BloomBench: A Bilingual Multimodal Benchmark for Cognitively Informed Evaluation of Vision-Language Models", "type": "paper", "country": "QA", "org": "QCRI", "license": "unknown", "modality": "multimodal", "tasks": [ "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2606.05531", "github": "https://github.com/qcri/Almieyar-Oryx-BloomBench" }, "year": 2026, "venue": "arXiv 2026", "notes": "We introduce BloomBench, part of the Almieyar benchmarking series, the first cognitively human-grounded, bilingual.", "metrics": null }, { "id": "almieyar-a-culturally-grounded-benchmark-for-multi-dialect-arabic-spee", "name": "Almieyar: A Culturally Grounded Benchmark for Multi-Dialect Arabic Speech Recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "evaluation", "speech", "vision" ], "links": { "paper": "https://arxiv.org/abs/2609.35564" }, "dialects": [ "lev", "msa", "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "We introduce ALMIEYAR, a culturally grounded ASR benchmark covering 17 Arabic dialects across six families.", "metrics": null }, { "id": "an-end-to-end-hybrid-framework-for-rumour-detection-in-low-resources-a", "name": "An End-to-End Hybrid Framework for Rumour Detection in Low-Resources Algerian Dialect", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2606.13411" }, "dialects": [ "magh" ], "year": 2026, "venue": "arXiv 2026", "notes": "This paper presents an end-to-end rumour detection hybrid framework for Algerian dialect social media content.", "metrics": null }, { "id": "arabdiscrim-a-decade-long-arabic-facebook-corpus-on-racism-and-discrim", "name": "ArabDiscrim: A Decade-Long Arabic Facebook Corpus on Racism and Discrimination", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "vision" ], "links": { "paper": "https://arxiv.org/abs/2605.22081" }, "year": 2026, "venue": "arXiv 2026", "notes": "We present ArabDiscrim, a decade-long lexical resource and corpus of 293K public Arabic Facebook posts (2014--2024) discussing racism and discrimination.", "metrics": null }, { "id": "arabic-prompts-with-english-tools-a-benchmark", "name": "Arabic Prompts with English Tools: A Benchmark", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2601.05101" }, "year": 2026, "venue": "arXiv 2026", "notes": "Large Language Models (LLMs) are now integral to numerous industries, increasingly serving as the core reasoning engine for autonomous agents.", "metrics": null }, { "id": "arabicdialecthub-a-cross-dialectal-arabic-learning-resource-and-platfo", "name": "ArabicDialectHub: A Cross-Dialectal Arabic Learning Resource and Platform", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation" ], "links": { "paper": "https://arxiv.org/abs/2601.22987", "website": "https://arabic-dialect-hub.netlify.app" }, "dialects": [ "gulf", "lev", "magh", "msa" ], "year": 2026, "venue": "arXiv 2026", "notes": "We present ArabicDialectHub, a cross-dialectal Arabic learning resource comprising 552 phrases across six varieties.", "metrics": null }, { "id": "arabicdialectsafety-a-dialect-aware-benchmark-for-arabic-content-safet", "name": "ArabicDialectSafety: A Dialect-Aware Benchmark for Arabic Content Safety Classification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "safety", "dialects", "classification" ], "links": { "paper": "https://arxiv.org/abs/2608.01291" }, "dialects": [ "msa", "egy", "lev", "magh" ], "year": 2026, "venue": "arXiv 2026", "notes": "ArabicDialectSafety: 25,071 human-curated safety prompts across MSA and four dialect groups, with seven harm categories and a dual-task evaluation.", "metrics": null }, { "id": "aradetox-a-multi-dialect-arabic-detoxification-dataset", "name": "AraDetox: A Multi-Dialect Arabic Detoxification Dataset", "type": "paper", "country": "INTL", "org": "arabicnlp-uk", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2608.22894", "github": "https://github.com/ArabicNLP-UK/AraDetox" }, "dialects": [ "egy", "gulf", "lev", "msa", "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "We introduce AraDetox, a multi-dialect Arabic detoxification dataset comprising 10,500 harmful social-media posts and 84,000 detoxified rewrites.", "metrics": null }, { "id": "arafa-an-llm-generated-arabic-fact-checking-dataset", "name": "ARAFA: An LLM-Generated Arabic Fact-Checking Dataset", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2609.25833" }, "dialects": [ "msa" ], "year": 2026, "venue": "arXiv 2026", "notes": "We introduce Arafa, a new large-scale dataset for fact-checking in Modern Standard Arabic, constructed through an automated framework.", "metrics": null }, { "id": "aragenre-2026-a-hierarchical-definition-guided-arabic-genre-classifica", "name": "AraGenre 2026: A Hierarchical Definition-Guided Arabic Genre Classification Shared Task", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2609.27387" }, "dialects": [ "classical", "msa", "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "AraGenre is a shared task on hierarchical, definition-guided Arabic genre classification, motivated by the limited availability of annotated data in Arabic.", "metrics": null }, { "id": "arahopecorpus-annotation-guidelines-and-dataset-for-hope-speech-in-ara", "name": "AraHopeCorpus: Annotation Guidelines and Dataset for Hope Speech in Arabic Social Media Crisis Discourse", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "chat", "sentiment", "speech" ], "links": { "paper": "https://arxiv.org/abs/2605.23325" }, "year": 2026, "venue": "arXiv 2026", "notes": "This paper introduces AraHopeCorpus, the first annotated dataset of Arabic hope speech collected from ten thousand YouTube comments related to the war.", "metrics": null }, { "id": "arams-28k-the-largest-publicly-released-line-level-dataset-of-historic", "name": "AraMS-28k: The Largest Publicly Released Line-Level Dataset of Historical Arabic Manuscripts with Margin and Insertion-Anchor Annotations", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "asr", "ocr", "diacritization", "vision" ], "links": { "paper": "https://arxiv.org/abs/2608.26921" }, "dialects": [ "magh" ], "year": 2026, "venue": "arXiv 2026", "notes": "We introduce AraMS-28k, the largest publicly released line-level dataset of genuine historical Arabic manuscripts, comprising 14 books, 3,043 pages.", "metrics": null }, { "id": "arcade-a-city-scale-corpus-for-fine-grained-arabic-dialect-tagging", "name": "ARCADE: A City-Scale Corpus for Fine-Grained Arabic Dialect Tagging", "type": "paper", "country": "SA", "org": "riotu-lab", "license": "cc-by-4.0", "modality": "speech", "tasks": [ "dialect-id", "sentiment", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2601.02209", "hf": "https://huggingface.co/datasets/riotu-lab/ARCADE-full" }, "dialects": [ "msa", "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "We present ARCADE (Arabic Radio Corpus for Audio Dialect Evaluation), the first Arabic speech dataset designed explicitly with city-level dialect granularity.", "metrics": { "downloads": 117, "likes": 5, "lastModified": "2026-01-12" } }, { "id": "are-arabic-benchmarks-reliable-qimma-s-quality-first-approach-to-llm-e", "name": "Are Arabic Benchmarks Reliable? QIMMA's Quality-First Approach to LLM Evaluation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2604.03395" }, "year": 2026, "venue": "arXiv 2026", "notes": "We present QIMMA, a quality-assured Arabic LLM leaderboard that places systematic benchmark validation at its core.", "metrics": null }, { "id": "arpomeme-an-annotated-arabic-multimodal-dataset-for-political-ideology", "name": "ArPoMeme: An Annotated Arabic Multimodal Dataset for Political Ideology and Polarization", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "vision" ], "links": { "paper": "https://arxiv.org/abs/2605.20967" }, "year": 2026, "venue": "arXiv 2026", "notes": "This paper presents ArPoMeme, a large-scale dataset of approximately 7,300 Arabic political memes categorized by ideological orientation.", "metrics": null }, { "id": "audience-engagement-with-arabic-women-s-social-empowerment-and-wellbei", "name": "Audience Engagement with Arabic Women's Social Empowerment and Wellbeing: A Decadal Corpus", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://arxiv.org/abs/2605.22204" }, "dialects": [ "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "This paper presents the Arabic Women and Society Corpus, a ten year collection of 252,487 public Arabic Facebook posts related to women's empowerment.", "metrics": null }, { "id": "babeljudge-measuring-llm-as-a-judge-reliability-across-languages-and-a", "name": "BabelJudge: Measuring LLM-as-a-Judge Reliability Across Languages and Agent Trajectories", "type": "paper", "country": "INTL", "org": "shreyaskc", "license": "unknown", "modality": "text", "tasks": [ "chat", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2606.22329", "github": "https://github.com/Shreyaskc/BabelJudge" }, "year": 2026, "venue": "arXiv 2026", "notes": "We introduce BabelJudge, an open-source benchmark and reliability audit framework that measures all four failure modes.", "metrics": null }, { "id": "benchmarking-arabic-russian-machine-translation-a-comparison-of-fine-t", "name": "Benchmarking Arabic--Russian Machine Translation: A Comparison of Fine-tuned NMT and Few-shot LLMs under Rich Morphology and Low Lexical Overlap", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2609.29559" }, "year": 2026, "venue": "arXiv 2026", "notes": "Arabic-Russian machine translation (MT) remains under-explored due to the rich morphology of Arabic and low lexical overlap between the two languages.", "metrics": null }, { "id": "benchmarking-commercial-asr-systems-on-code-switching-speech-arabic-pe", "name": "Benchmarking Commercial ASR Systems on Code-Switching Speech: Arabic, Persian, and German", "type": "paper", "country": "INTL", "org": "perle-ai", "license": "mit", "modality": "speech", "tasks": [ "asr", "embedding", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2605.19069", "hf": "https://huggingface.co/datasets/Perle-ai/ASR_Code_Switch" }, "dialects": [ "egy", "gulf" ], "year": 2026, "venue": "arXiv 2026", "notes": "We present a benchmark evaluating five commercial ASR providers across four language pairs: Egyptian Arabic--English, Saudi Arabic.", "metrics": { "downloads": 923, "likes": 12, "lastModified": "2026-05-22" } }, { "id": "benchmarking-frontier-llms-on-arabic-cultural-and-sociolinguistic-know", "name": "Benchmarking Frontier LLMs on Arabic Cultural and Sociolinguistic Knowledge: A Cross-Evaluation Framework with Human SME Ground Truth", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2607.00139" }, "dialects": [ "egy", "iraqi" ], "year": 2026, "venue": "arXiv 2026", "notes": "The cost of human expert evaluation is a principal bottleneck to deploying language models in specialized, high-stakes domains.", "metrics": null }, { "id": "beyond-fluency-a-rubric-based-benchmark-for-evaluating-saudi-dialect-a", "name": "Beyond Fluency: A Rubric-Based Benchmark for Evaluating Saudi Dialect and Cultural Competence in Large Language Models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2608.29990" }, "dialects": [ "gulf", "msa" ], "year": 2026, "venue": "arXiv 2026", "notes": "We present a rubric-based benchmark for the Saudi dialect, comprising 31 expert-authored prompts spanning idiomatic, pragmatic, lexical.", "metrics": null }, { "id": "bolbosh-script-aware-flow-matching-for-kashmiri-text-to-speech", "name": "Bolbosh: Script-Aware Flow Matching for Kashmiri Text-to-Speech", "type": "paper", "country": "INTL", "org": "gaash-lab", "license": "unknown", "modality": "speech", "tasks": [ "tts", "diacritization", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2603.07513", "github": "https://github.com/gaash-lab/Bolbosh" }, "year": 2026, "venue": "arXiv 2026", "notes": "In this work, we present the first dedicated open-source neural TTS system designed for Kashmiri.", "metrics": null }, { "id": "bridging-scientific-heritage-an-arabic-russian-parallel-corpus-and-llm", "name": "Bridging Scientific Heritage: An Arabic--Russian Parallel Corpus and LLM Benchmark for Sustainable Knowledge Transfer", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "chat", "translation", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2606.30943" }, "year": 2026, "venue": "arXiv 2026", "notes": "We present a benchmark for Arabic--Russian scientific translation.", "metrics": null }, { "id": "bulbul-a-dataset-for-dialectal-arabic-speech-recognition", "name": "Bulbul: A Dataset for Dialectal Arabic Speech Recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2608.21950" }, "dialects": [ "classical", "msa", "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "We present BULBUL, a multi-dialect Arabic ASR dataset collected from 275 speakers in 11 Arab countries.", "metrics": null }, { "id": "candle-ctc-based-arabic-noisy-character-deduplication-using-a-lightwei", "name": "CANDLE: CTC-based Arabic Noisy-character Deduplication using a Lightweight Encoder", "type": "paper", "country": "INTL", "org": "Abjad AI", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2606.24758", "github": "https://github.com/abjadai/candle" }, "year": 2026, "venue": "arXiv 2026", "notes": "We present CANDLE, a lightweight system for character-level Arabic noise deduplication that addresses this challenge without relying on handcrafted rules.", "metrics": null }, { "id": "cardamom-a-micro-dialectal-arabic-speech-dataset-for-asr", "name": "CARDAMOM: A Micro-Dialectal Arabic Speech Dataset for ASR", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "dialects", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2609.34481" }, "dialects": [ "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "Cardamom: about 40 hours of transcribed YouTube speech across 21 micro-dialects for fine-grained Arabic ASR evaluation.", "metrics": null }, { "id": "cohesion-6k-an-arabic-dataset-for-analyzing-social-cohesion-and-confli", "name": "Cohesion-6K: An Arabic Dataset for Analyzing Social Cohesion and Conflict in Online Discourse", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "discourse", "social-media", "classification" ], "links": { "paper": "https://arxiv.org/abs/2605.22447" }, "year": 2026, "venue": "arXiv 2026", "notes": "Cohesion-6K: 6,000 Arabic Facebook posts on the Israeli occupation of Palestine, labelled on a conflict-to-cohesion discourse scale.", "metrics": null }, { "id": "cross-lingual-learning-within-arabic-script-for-low-resource-htr", "name": "Cross-Lingual Learning within Arabic Script for Low-Resource HTR", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "ocr", "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2605.02089" }, "year": 2026, "venue": "arXiv 2026", "notes": "Handwritten Text Recognition (HTR) with limited labeled data remains a challenging problem, particularly for Arabic-script languages.", "metrics": null }, { "id": "crosshallu-do-hallucination-signals-generalize-across-languages-and-do", "name": "CrossHallu: Do Hallucination Signals Generalize Across Languages and Domains in Large Language Model's Internals?", "type": "paper", "country": "INTL", "org": "aishaalansari57", "license": "unknown", "modality": "text", "tasks": [ "translation", "qa", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2607.04029", "github": "https://github.com/aishaalansari57/CrossHal" }, "year": 2026, "venue": "arXiv 2026", "notes": "We present CrossHallu, the first study to evaluate the cross-lingual and cross-domain generalization of hallucination detection.", "metrics": null }, { "id": "cultural-benchmarking-of-llms-in-standard-and-dialectal-arabic-dialogu", "name": "Cultural Benchmarking of LLMs in Standard and Dialectal Arabic Dialogues", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "chat", "translation", "qa", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2605.00119" }, "dialects": [ "msa", "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "We introduce ArabCulture-Dialogue, a culturally grounded conversational dataset covering 13 Arabic-speaking countries.", "metrics": null }, { "id": "curriculum-learning-and-pseudo-labeling-improve-the-generalization-of", "name": "Curriculum Learning and Pseudo-Labeling Improve the Generalization of Multi-Label Arabic Dialect Identification Models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2602.12937", "website": "https://mohamedalaa9.github.io/lahjatbert" }, "dialects": [ "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "We construct a multi-label dataset by generating automatic multi-label annotations using GPT-4o and binary dialect acceptability classifiers.", "metrics": null }, { "id": "cv-18-ner-augmented-common-voice-for-named-entity-recognition-from-ara", "name": "CV-18 NER: Augmented Common Voice for Named Entity Recognition from Arabic Speech", "type": "paper", "country": "INTL", "org": "elyadata", "license": "cc-by-nc-4.0", "modality": "speech", "tasks": [ "asr", "ner", "pretraining", "evaluation", "speech", "vision" ], "links": { "paper": "https://arxiv.org/abs/2604.02209", "hf": "https://huggingface.co/datasets/Elyadata/CV18-NER" }, "year": 2026, "venue": "arXiv 2026", "notes": "We introduce CV-18 NER, the first publicly available dataset for NER from Arabic speech, created by augmenting the Arabic Common Voice 18 corpus.", "metrics": { "downloads": 99, "likes": 1, "lastModified": "2026-05-19" } }, { "id": "do-llms-fabricate-legal-citations-a-bilingual-benchmark-on-saudi-data", "name": "Do LLMs Fabricate Legal Citations? A Bilingual Benchmark on Saudi Data Protection Law and the GDPR", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2607.11127" }, "dialects": [ "gulf" ], "year": 2026, "venue": "arXiv 2026", "notes": "We introduce a bilingual benchmark of 120 questions probing whether freely accessible LLMs fabricate article citations for two data-protection instruments.", "metrics": null }, { "id": "e-conan-entailment-contradition-and-neutral-benchmarks-arabic-textual", "name": "E-CONAN (Entailment, CONtradition And Neutral) Benchmarks: Arabic Textual Entailment and Natural Inference Datasets", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2609.11334" }, "year": 2026, "venue": "arXiv 2026", "notes": "This paper introduces E-CONAN benchmarks that are composed of sentences pairs from various sources: (1) automatically-translated pairs.", "metrics": null }, { "id": "e-conan-entailment-contradition-and-neutral-diagnostics-dataset-invest", "name": "E-CONAN (Entailment, CONtradition And Neutral) Diagnostics Dataset Investigating Linguistic Phenomena in Arabic Natural Language Understanding", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2609.33530" }, "year": 2026, "venue": "arXiv 2026", "notes": "To overcome this gap, we propose an initial hierarchy for Cross-Lingual NLU error analysis.", "metrics": null }, { "id": "edrac-benchmarking-arabic-dialect-reading-comprehension", "name": "EDRAC: Benchmarking Arabic Dialect Reading Comprehension", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "qa", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2609.01113" }, "dialects": [ "egy", "gulf", "lev", "magh", "msa", "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "We introduce EDRAC, the first large-scale benchmark for dialectal Arabic machine reading comprehension (MRC) and generative QA, covering five major dialects.", "metrics": null }, { "id": "efficient-multilingual-neural-machine-translation-via-corpus-driven-vo", "name": "Efficient Multilingual Neural Machine Translation via Corpus-Driven Vocabulary Pruning: An English-Arabic Case Study", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "translation", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2608.03480" }, "year": 2026, "venue": "arXiv 2026", "notes": "We propose in this paper a general optimization framework that combines a vocabulary pruning method with a targeted fine-tuning protocol for MNMT models.", "metrics": null }, { "id": "fanar-sadiq-a-multi-agent-architecture-for-grounded-islamic-qa", "name": "Fanar-Sadiq: A Multi-Agent Architecture for Grounded Islamic QA", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "qa", "evaluation", "quran", "hadith" ], "links": { "paper": "https://arxiv.org/abs/2603.08501", "website": "https://api.fanar.qa/docs" }, "dialects": [ "classical" ], "year": 2026, "venue": "arXiv 2026", "notes": "To address these challenges, we present Fanar-Sadiq, a bilingual Arabic-English Islamic QA system built on a multi-agent, tool-augmented architecture.", "metrics": null }, { "id": "grounding-arabic-llms-in-the-doha-historical-dictionary-retrieval-augm", "name": "Grounding Arabic LLMs in the Doha Historical Dictionary: Retrieval-Augmented Understanding of Quran and Hadith", "type": "paper", "country": "QA", "org": "somayaeltanbouly", "license": "unknown", "modality": "text", "tasks": [ "rag", "quran", "lexicography" ], "links": { "paper": "https://arxiv.org/abs/2603.23972", "github": "https://github.com/somayaeltanbouly/Doha-Dictionary-RAG" }, "dialects": [ "classical" ], "year": 2026, "venue": "arXiv 2026", "notes": "RAG framework grounded in the Doha Historical Dictionary of Arabic that lifts Fanar and ALLaM above 85% on Quran and Hadith questions.", "metrics": null }, { "id": "habibi-laying-the-open-source-foundation-of-unified-dialectal-arabic-s", "name": "Habibi: Laying the Open-Source Foundation of Unified-Dialectal Arabic Speech Synthesis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "tts", "diacritization", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2601.13802", "website": "https://SWivid.github.io/Habibi" }, "dialects": [ "msa", "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "We present Habibi, a unified-dialectal Arabic TTS framework that addresses all three.", "metrics": null }, { "id": "hallutruthqa-4k-a-fine-grained-corpus-and-annotation-process-for-arabi", "name": "HalluTruthQA-4K: A Fine-Grained Corpus and Annotation Process for Arabic Hallucination Detection and Truth Verification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2608.03966" }, "year": 2026, "venue": "arXiv 2026", "notes": "We present HalluTruthQA-4K, an expanded version of the HalluTruthQA resource containing 4,000 expert-curated Arabic question-answering instances.", "metrics": null }, { "id": "hallutruthqa-a-fine-grained-benchmark-for-hallucination-detection-loca", "name": "HalluTruthQA: A Fine-Grained Benchmark for Hallucination Detection, Localization, and Explanation in Arabic Question Answering", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2607.20219" }, "year": 2026, "venue": "arXiv 2026", "notes": "We introduce HalluTruthQA, a fine-grained benchmark for hallucination evaluation in Arabic question answering.", "metrics": null }, { "id": "halluverse-m-3-a-multitask-multilingual-benchmark-for-hallucination-in", "name": "Halluverse-M^3: A multitask multilingual benchmark for hallucination in LLMs", "type": "paper", "country": "INTL", "org": "sabdalja", "license": "unknown", "modality": "text", "tasks": [ "qa", "summarization", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2602.06920", "hf": "https://huggingface.co/datasets/sabdalja/HalluVerse-M3" }, "year": 2026, "venue": "arXiv 2026", "notes": "We introduce Halluverse-M^3, a dataset designed to enable systematic analysis of hallucinations across multiple languages, multiple generation tasks.", "metrics": { "downloads": 72, "likes": 2, "lastModified": "2026-05-04" } }, { "id": "instruction-guided-poetry-generation-in-arabic-and-its-dialects", "name": "Instruction-Guided Poetry Generation in Arabic and Its Dialects", "type": "paper", "country": "AE", "org": "MBZUAI", "license": "unknown", "modality": "text", "tasks": [ "evaluation", "poetry" ], "links": { "paper": "https://arxiv.org/abs/2604.27766", "github": "https://github.com/mbzuai-nlp/instructpoet-ar" }, "dialects": [ "msa", "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "Specifically, we present a large-scale, carefully curated instruction-based dataset in Modern Standard Arabic (MSA) and various Arabic dialects.", "metrics": null }, { "id": "islamicmmlu-a-benchmark-for-evaluating-llms-on-islamic-knowledge", "name": "IslamicMMLU: A Benchmark for Evaluating LLMs on Islamic Knowledge", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa", "evaluation", "quran", "hadith" ], "links": { "paper": "https://arxiv.org/abs/2603.23750" }, "dialects": [ "classical" ], "year": 2026, "venue": "arXiv 2026", "notes": "We introduce IslamicMMLU, a benchmark of 10,013 multiple-choice questions spanning three tracks: Quran (2,013 questions), Hadith (4,000 questions), and Fiqh.", "metrics": null }, { "id": "jobarabi-an-arabic-corpus-and-analysis-of-job-announcements-from-socia", "name": "JobArabi: An Arabic Corpus and Analysis of Job Announcements from Social Media", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://arxiv.org/abs/2605.20960" }, "year": 2026, "venue": "arXiv 2026", "notes": "This paper introduces JobArabi, a large-scale corpus of Arabic job announcements collected from social media between January 2024 and October 2025.", "metrics": null }, { "id": "linear-semantic-segmentation-for-low-resource-spoken-dialects", "name": "Linear Semantic Segmentation for Low-Resource Spoken Dialects", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "chat", "asr", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2605.06276" }, "dialects": [ "msa", "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "We introduce a new multi-genre benchmark (more than 1000 samples) for semantic segmentation in conversational Arabic, focusing on dialectal discourse.", "metrics": null }, { "id": "lqm-linguistically-motivated-multidimensional-quality-metrics-for-mach", "name": "LQM: Linguistically Motivated Multidimensional Quality Metrics for Machine Translation", "type": "paper", "country": "INTL", "org": "UBC-NLP", "license": "unknown", "modality": "text", "tasks": [ "chat", "translation", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2604.18490", "github": "https://github.com/UBC-NLP/LQM_MT" }, "dialects": [ "egy", "gulf", "lev", "magh", "yemeni", "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "However, they often fail to capture dialect- and culture-specific errors in diglossic languages (e.g., Arabic).", "metrics": null }, { "id": "macaron-controlled-human-written-benchmark-for-multilingual-and-multic", "name": "Macaron: Controlled, Human-Written Benchmark for Multilingual and Multicultural Reasoning via Template-Filling", "type": "paper", "country": "INTL", "org": "alaaahmed2444", "license": "unknown", "modality": "text", "tasks": [ "translation", "qa", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2602.10732", "hf": "https://huggingface.co/datasets/AlaaAhmed2444/Macaron" }, "dialects": [ "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "We propose Macaron, a template-first benchmark that factorizes reasoning type and cultural aspect across question languages.", "metrics": { "downloads": 98, "likes": 2, "lastModified": "2026-02-14" } }, { "id": "mawarith-a-dataset-and-benchmark-for-legal-inheritance-reasoning-with", "name": "MAWARITH: A Dataset and Benchmark for Legal Inheritance Reasoning with LLMs", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2603.07539", "website": "https://gitlab.com/nlpresearcher/mawarith" }, "year": 2026, "venue": "arXiv 2026", "notes": "MAWARITH: 12,500 annotated Arabic Islamic inheritance cases for training and evaluating step-by-step inheritance reasoning.", "metrics": null }, { "id": "mawqif-xt-an-arabic-benchmark-dataset-for-cross-target-stance-detectio", "name": "Mawqif-XT: An Arabic Benchmark Dataset for Cross-Target Stance Detection", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2608.09539" }, "year": 2026, "venue": "arXiv 2026", "notes": "This paper presents the Mawqif-XT, consisting of 996 manually annotated Arabic tweets collected from three public targets: Women Driving, E-Cars.", "metrics": null }, { "id": "medarabench-large-scale-arabic-medical-question-answering-dataset-and", "name": "MedAraBench: Large-Scale Arabic Medical Question Answering Dataset and Benchmark", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2602.01714" }, "year": 2026, "venue": "arXiv 2026", "notes": "In this paper, we introduce MedAraBench, a large-scale dataset consisting of Arabic multiple-choice question-answer pairs across various medical specialties.", "metrics": null }, { "id": "mederrbench-a-fine-grained-multilingual-benchmark-for-medical-error-de", "name": "MedErrBench: A Fine-Grained Multilingual Benchmark for Medical Error Detection and Correction with Clinical Expert Annotations", "type": "paper", "country": "INTL", "org": "congboma", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2602.05692", "github": "https://github.com/congboma/MedErrBench" }, "year": 2026, "venue": "arXiv 2026", "notes": "We introduce MedErrBench, the first multilingual benchmark for error detection, localization.", "metrics": null }, { "id": "mizan-a-national-benchmark-for-evaluating-large-language-models-on-ira", "name": "Mizan: A National Benchmark for Evaluating Large Language Models on Iraqi Arabic and the Iraqi Civic Context", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialects", "evaluation", "safety" ], "links": { "paper": "https://arxiv.org/abs/2609.13980" }, "dialects": [ "iraqi", "msa" ], "year": 2026, "venue": "arXiv 2026", "notes": "Mizan: national benchmark of 340 items evaluating LLMs on Iraqi Arabic and Iraqi civic context across six axes plus an MSA baseline track.", "metrics": null }, { "id": "mudawansn-a-gold-standard-wolof-arabic-parallel-corpus-for-machine-tra", "name": "MudawanSn: A Gold-Standard Wolof-Arabic Parallel Corpus for Machine Translation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2609.17539" }, "dialects": [ "msa" ], "year": 2026, "venue": "arXiv 2026", "notes": "We present MudawanSn, a gold-standard resource of 1,271 sentence-aligned pairs manually translated from Wolof into Modern Standard Arabic (MSA).", "metrics": null }, { "id": "multi-task-instruction-tuning-via-data-scheduling-for-low-resource-ara", "name": "Multi-Task Instruction Tuning via Data Scheduling for Low-Resource Arabic SpeechLLMs", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "dialect-id", "sentiment", "summarization", "instruction-tuning", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2601.12494" }, "year": 2026, "venue": "arXiv 2026", "notes": "We present a controlled study of multi-task instruction tuning for an Arabic-centric audio LLM across generative tasks.", "metrics": null }, { "id": "multilingual-multi-label-emotion-classification-at-scale-with-syntheti", "name": "Multilingual Multi-Label Emotion Classification at Scale with Synthetic Data", "type": "paper", "country": "INTL", "org": "tabularisai", "license": "cc-by-nc-4.0", "modality": "text", "tasks": [ "sentiment", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2604.12633", "hf": "https://huggingface.co/tabularisai/multilingual-emotion-classification" }, "year": 2026, "venue": "arXiv 2026", "notes": "Emotion classification in multilingual settings remains constrained by the scarcity of annotated data.", "metrics": { "downloads": 2762, "likes": 27, "lastModified": "2026-07-31" } }, { "id": "murad-a-large-scale-multi-domain-unified-reverse-arabic-dictionary-dat", "name": "MURAD: A Large-Scale Multi-Domain Unified Reverse Arabic Dictionary Dataset", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding" ], "links": { "paper": "https://arxiv.org/abs/2601.21512" }, "year": 2026, "venue": "arXiv 2026", "notes": "We present MURAD (Multi-domain Unified Reverse Arabic Dictionary), an open lexical dataset with 96,243 word-definition pairs.", "metrics": null }, { "id": "nadi-2026-the-second-multidialectal-arabic-speech-processing-shared-ta", "name": "NADI 2026: The Second Multidialectal Arabic Speech Processing Shared Task", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "asr", "tts", "translation", "dialect-id", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2609.27086" }, "dialects": [ "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "NADI 2026 is the seventh edition of the Nuanced Arabic Dialect Identification.", "metrics": null }, { "id": "obscuring-data-contamination-through-translation-evidence-from-arabic", "name": "Obscuring Data Contamination Through Translation: Evidence from Arabic Corpora", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2601.14994" }, "year": 2026, "venue": "arXiv 2026", "notes": "We propose Translation-Aware Contamination Detection, which identifies contamination by comparing signals.", "metrics": null }, { "id": "once-correct-still-wrong-counterfactual-hallucination-in-multilingual", "name": "Once Correct, Still Wrong: Counterfactual Hallucination in Multilingual Vision-Language Models", "type": "paper", "country": "QA", "org": "QCRI", "license": "cc-by-nc-sa-4.0", "modality": "multimodal", "tasks": [ "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2602.05437", "hf": "https://huggingface.co/datasets/QCRI/M2CQA" }, "dialects": [ "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "We introduce M$^2$CQA, a culturally grounded multimodal benchmark built from images spanning 17 MENA countries.", "metrics": { "downloads": 662, "likes": 2, "lastModified": "2026-06-12" } }, { "id": "qias-2026-overview-of-the-shared-task-on-islamic-inheritance-reasoning", "name": "QIAS 2026: Overview of the Shared Task on Islamic Inheritance Reasoning", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "qa", "summarization", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2606.13756" }, "year": 2026, "venue": "arXiv 2026", "notes": "This paper presents a comprehensive overview of the QIAS 2026 shared task, organized as part of the OSACT7 Workshop and co-located with LREC 2026.", "metrics": null }, { "id": "quran-md-a-fine-grained-multilingual-multimodal-dataset-of-the-quran", "name": "Quran-MD: A Fine-Grained Multilingual Multimodal Dataset of the Quran", "type": "paper", "country": "INTL", "org": "buraaq", "license": "unknown", "modality": "multimodal", "tasks": [ "asr", "tts", "embedding", "translation", "speech", "quran" ], "links": { "paper": "https://arxiv.org/abs/2601.17880", "hf": "https://huggingface.co/datasets/Buraaq/quran-audio-text-dataset" }, "dialects": [ "classical" ], "year": 2026, "venue": "arXiv 2026", "notes": "We present Quran MD, a comprehensive multimodal dataset of the Quran that integrates textual, linguistic, and audio dimensions at the verse and word levels.", "metrics": { "downloads": 132, "likes": 16, "lastModified": "2026-05-24" } }, { "id": "quranicmmlu-a-cognitively-aware-benchmark-for-evaluating-generative-ai", "name": "QuranicMMLU: A Cognitively-Aware Benchmark for Evaluating Generative AI Solutions on Quranic Linguistic Knowledge", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "qa", "evaluation", "quran" ], "links": { "paper": "https://arxiv.org/abs/2609.22038" }, "dialects": [ "classical" ], "year": 2026, "venue": "arXiv 2026", "notes": "We introduce QuranicMMLU, a benchmark for evaluating generative AI on Quranic Arabic across multiple dimensions of linguistic complexity.", "metrics": null }, { "id": "ramsa-a-large-sociolinguistically-rich-emirati-arabic-speech-corpus-fo", "name": "Ramsa: A Large Sociolinguistically Rich Emirati Arabic Speech Corpus for ASR and TTS", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "tts", "evaluation", "speech", "vision" ], "links": { "paper": "https://arxiv.org/abs/2603.08125" }, "dialects": [ "gulf" ], "year": 2026, "venue": "arXiv 2026", "notes": "Ramsa is a developing 41-hour speech corpus of Emirati Arabic designed to support sociolinguistic research and low-resource language technologies.", "metrics": null }, { "id": "rightnow-arabic-0-5b-turbo-an-open-sub-1b-arabic-language-model-via-vo", "name": "RightNow-Arabic-0.5B-Turbo: An Open Sub-1B Arabic Language Model via Vocabulary Injection and Edge-First Deployment", "type": "paper", "country": "INTL", "org": "rightnowai", "license": "apache-2.0", "modality": "text", "tasks": [ "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2605.28827", "hf": "https://huggingface.co/RightNowAI/RightNow-Arabic-0.5B-Turbo" }, "year": 2026, "venue": "arXiv 2026", "notes": "We present RightNow-Arabic-0.5B-Turbo, a 518M-parameter Arabic-specialized decoder LLM built on Qwen2.5-0.5B.", "metrics": { "downloads": 513, "likes": 6, "lastModified": "2026-04-10" } }, { "id": "sahm-a-benchmark-for-arabic-financial-and-shari-ah-compliant-reasoning", "name": "SAHM: A Benchmark for Arabic Financial and Shari'ah-Compliant Reasoning", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "qa", "summarization", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2604.19098" }, "dialects": [ "gulf" ], "year": 2026, "venue": "arXiv 2026", "notes": "We introduce Sahm, the first Arabic financial benchmark spanning seven tasks.", "metrics": null }, { "id": "shams-an-audio-grounded-pronunciation-benchmark-for-levantine-arabic", "name": "SHAMS: An Audio-Grounded Pronunciation Benchmark for Levantine Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "pronunciation", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2610.01427", "website": "https://shams-nlp.github.io" }, "dialects": [ "lev" ], "year": 2026, "venue": "arXiv 2026", "notes": "SHAMS: 1,300-utterance benchmark for Levantine Arabic speech, evaluating pronunciation under opaque, non-standardized orthography.", "metrics": null }, { "id": "syrisign-a-parallel-corpus-for-arabic-text-to-syrian-arabic-sign-langu", "name": "SyriSign: A Parallel Corpus for Arabic Text to Syrian Arabic Sign Language Translation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "embedding", "translation", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2603.29219" }, "dialects": [ "lev" ], "year": 2026, "venue": "arXiv 2026", "notes": "To overcome this gap, we introduce SyriSign, a dataset comprising 1500 video samples across 150 unique lexical signs.", "metrics": null }, { "id": "tarab-a-multi-dialect-corpus-of-arabic-lyrics-and-poetry", "name": "Tarab: A Multi-Dialect Corpus of Arabic Lyrics and Poetry", "type": "paper", "country": "INTL", "org": "drelhaj", "license": "cc-by-4.0", "modality": "text", "tasks": [ "poetry" ], "links": { "paper": "https://arxiv.org/abs/2603.16601", "hf": "https://huggingface.co/datasets/drelhaj/Tarab" }, "dialects": [ "egy", "gulf", "lev", "magh", "iraqi", "sudanese", "classical", "msa", "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "We introduce the Tarab Corpus, a large-scale cultural and linguistic resource that brings together Arabic song lyrics.", "metrics": { "downloads": 181, "likes": 4, "lastModified": "2026-02-25" } }, { "id": "terminal-bench-lilt-multilingual-agentic-coding-benchmark-grounded-in", "name": "Terminal-Bench-LILT: Multilingual Agentic Coding Benchmark Grounded in Language, Region, and Culture", "type": "paper", "country": "INTL", "org": "lilt", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2608.28641", "github": "https://github.com/lilt/terminal-bench-lilt" }, "year": 2026, "venue": "arXiv 2026", "notes": "We present Terminal-Bench-LILT, a suite of 300 authentic coding tasks in ten languages.", "metrics": null }, { "id": "the-generator-eraser-paradox-community-guidelines-for-responsible-llm", "name": "The Generator-Eraser Paradox: Community Guidelines for Responsible LLM-Assisted Dialect Resource Creation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding" ], "links": { "paper": "https://arxiv.org/abs/2606.06004" }, "dialects": [ "mixed" ], "year": 2026, "venue": "arXiv 2026", "notes": "Dialect resources occupy a unique position at the intersection of scientific description, cultural preservation, and computational infrastructure.", "metrics": null }, { "id": "toxirex-a-dataset-on-toxic-reasoning-in-context", "name": "ToxiREX: A Dataset on Toxic REasoning in ConteXt", "type": "paper", "country": "INTL", "org": "cltl", "license": "unknown", "modality": "text", "tasks": [ "chat", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2606.27981", "github": "https://github.com/cltl/toxirex" }, "year": 2026, "venue": "arXiv 2026", "notes": "We introduce a new, contextual, multilingual dataset called ToxiREX: Toxic REasoning in ConteXt.", "metrics": null }, { "id": "ttlab-at-alexandriax-2026-a-fine-tuned-surface-tagger-for-arabic-machi", "name": "TTLab at AlexandriaX-2026: A Fine-Tuned Surface Tagger for Arabic Machine-Translation Error-Span Detection and Classification", "type": "paper", "country": "INTL", "org": "entailab", "license": "unknown", "modality": "text", "tasks": [ "translation", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2609.29633", "github": "https://github.com/ENTAILab/arabic-dialectal-mt-error-span-detection" }, "year": 2026, "venue": "arXiv 2026", "notes": "We present TTLab's submission to the AlexandriaX-2026 Subtask3 on Arabic MT error span detection and classification.", "metrics": null }, { "id": "ttlab-at-daleel-2026-star-ar-sequence-tagging-for-argument-recognition", "name": "TTLab at Daleel 2026: STAR-Ar, Sequence Tagging for Argument Recognition in Arabic", "type": "paper", "country": "INTL", "org": "entailab", "license": "unknown", "modality": "text", "tasks": [ "embedding" ], "links": { "paper": "https://arxiv.org/abs/2609.39385", "github": "https://github.com/ENTAILab/daleel_2026_Arabic-Argumentative-Discourse-Mining" }, "year": 2026, "venue": "arXiv 2026", "notes": "Argument Mining (AM) is a critical NLP task that remains significantly under-resourced in Arabic.", "metrics": null }, { "id": "tutlait-v1-a-crowdsourced-moroccan-tamazight-speech-dataset-with-arabi", "name": "TutlAit v1: a crowdsourced Moroccan Tamazight speech dataset with Arabic transcriptions and regional accent labels", "type": "paper", "country": "MA", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "translation", "tamazight" ], "links": { "paper": "https://arxiv.org/abs/2609.38219" }, "year": 2026, "venue": "arXiv 2026", "notes": "TutlAit v1: crowdsourced Moroccan Tamazight speech paired with Arabic text and regional variety labels (Atlas, Souss, Rif).", "metrics": null }, { "id": "what-really-controls-temporal-reasoning-in-large-language-models-token", "name": "What Really Controls Temporal Reasoning in Large Language Models: Tokenisation or Representation of Time?", "type": "paper", "country": "INTL", "org": "gagan3012", "license": "unknown", "modality": "text", "tasks": [ "translation", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2603.19017", "github": "https://github.com/gagan3012/mtb" }, "year": 2026, "venue": "arXiv 2026", "notes": "We present MultiTempBench, a multilingual temporal reasoning benchmark spanning three tasks, date arithmetic, time zone conversion.", "metrics": null }, { "id": "when-do-vlms-help-arabic-manuscript-ocr-a-cross-dataset-study", "name": "When Do VLMs Help Arabic Manuscript OCR? A Cross-Dataset Study", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "ocr", "diacritization", "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2608.22366" }, "year": 2026, "venue": "arXiv 2026", "notes": "Vision-language models (VLMs) are increasingly being used for document understanding, yet their role in Arabic and Islamic manuscript recognition remains.", "metrics": null }, { "id": "yallamorph-a-benchmark-for-evaluating-arabic-morphological-generation", "name": "YallaMorph: A Benchmark for Evaluating Arabic Morphological Generation in Large Language Models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "diacritization", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2609.10153" }, "year": 2026, "venue": "arXiv 2026", "notes": "We introduce YallaMorph, a large-scale benchmark for Arabic morphological generation covering verbs, nouns, adjectives, their cliticized forms.", "metrics": null }, { "id": "ymir-a-new-benchmark-dataset-and-model-for-arabic-yemeni-music-genre-c", "name": "YMIR: A new Benchmark Dataset and Model for Arabic Yemeni Music Genre Classification Using Convolutional Neural Networks", "type": "paper", "country": "YE", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "music", "genre-classification" ], "links": { "paper": "https://arxiv.org/abs/2604.05011" }, "dialects": [ "yemeni" ], "year": 2026, "venue": "arXiv 2026", "notes": "YMIR: 1,475 audio clips of five Yemeni music genres labelled by experts, plus the YMCM CNN classifier.", "metrics": null }, { "id": "fanar-an-arabic-centric-multimodal-generative-ai-platform", "name": "Fanar: An Arabic-Centric Multimodal Generative AI Platform", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "multimodal" ], "links": { "paper": "https://arxiv.org/abs/2501.13944" }, "year": 2025, "venue": "arXiv.org", "citations": 91, "notes": "We present Fanar, a platform for Arabic-centric multimodal generative AI systems, that supports language, speech and image generation tasks.", "metrics": null }, { "id": "arabic-natural-language-processing-nlp-a-comprehensive-revie", "name": "Arabic Natural Language Processing (NLP): A Comprehensive Review of Challenges, Techniques, and Emerging Trends", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey" ], "links": { "paper": "https://doi.org/10.3390/computers14110497" }, "year": 2025, "venue": "De Computis", "citations": 31, "notes": "Arabic natural language processing (NLP) has garnered significant attention in recent years due to the growing demand for automated text and Arabic-based", "metrics": null }, { "id": "abstractive-text-summarization-in-arabic-like-script-using-m", "name": "Abstractive Text Summarization in Arabic-Like Script Using Multi-Encoder Architecture and Semantic Extraction Techniques", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "summarization" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2025.3575610" }, "year": 2025, "venue": "IEEE Access", "citations": 29, "notes": "In the field of Natural Language Processing (NLP), the task of text summarization plays a vital role in understanding textual content and producing concise", "metrics": null }, { "id": "evaluation-of-ai-generated-reading-comprehension-materials-f", "name": "Evaluation of AI-generated reading comprehension materials for Arabic language teaching", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://doi.org/10.1080/09588221.2025.2474037" }, "year": 2025, "venue": "Computer Assisted Language Learning", "citations": 29, "notes": "This study investigates the effectiveness of AI tools in generating Arabic reading comprehension materials, focusing on", "metrics": null }, { "id": "chinese-generative-ai-models-deepseek-and-qwen-rival-chatgpt", "name": "Chinese generative AI models (DeepSeek and Qwen) rival ChatGPT-4 in ophthalmology queries with excellent performance in Arabic and English", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://doi.org/10.52225/narra.v5i1.2371" }, "year": 2025, "venue": "Narra J", "citations": 27, "notes": "The rapid evolution of generative artificial intelligence (genAI) has ushered in a new era of digital medical consultations, with patients turning to AI-driven", "metrics": null }, { "id": "arabic-fake-news-detection-using-hybrid-contextual-features", "name": "Arabic fake news detection using hybrid contextual features", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification" ], "links": { "paper": "https://doi.org/10.11591/ijece.v15i1.pp836-845" }, "year": 2025, "venue": "International Journal of Electrical and Computer Engineering (IJECE)", "citations": 25, "notes": "Technology has advanced and social media users have grown dramatically in the last decade.", "metrics": null }, { "id": "towards-inclusive-arabic-llms-a-culturally-aligned-benchmark", "name": "Towards Inclusive Arabic LLMs: A Culturally Aligned Benchmark in Arabic Large Language Model Evaluation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining", "evaluation" ], "links": { "paper": "https://www.semanticscholar.org/paper/1ca6c715cd6664d25d8e9d5deda50f067359558b" }, "year": 2025, "venue": "COLING Workshops", "citations": 22, "metrics": null }, { "id": "a-scoping-review-of-arabic-natural-language-processing-for-m", "name": "A Scoping Review of Arabic Natural Language Processing for Mental Health", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey" ], "links": { "paper": "https://doi.org/10.3390/healthcare13090963" }, "year": 2025, "venue": "Healthcare", "citations": 21, "notes": "Mental health disorders represent a substantial global health concern, impacting millions and placing a significant burden on public health systems.", "metrics": null }, { "id": "command-r7b-arabic-a-small-enterprise-focused-multilingual-a", "name": "Command R7B Arabic: A Small, Enterprise Focused, Multilingual, and Culturally Aware Arabic LLM", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2503.14603" }, "year": 2025, "venue": "Proceedings of the Sixth Workshop on African Natural Language Processing (AfricaNLP 2025)", "citations": 21, "notes": "In this work, we present a data synthesis and refinement strategy to help address this problem, namely, by leveraging synthetic", "metrics": null }, { "id": "arabic-transliteration-of-borrowed-english-nouns-with-g-by-a", "name": "Arabic Transliteration of Borrowed English Nouns with /g/ by Artificial Intelligence (AI", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "preprocessing" ], "links": { "paper": "https://doi.org/10.32996/jcsts.2025.7.9.29" }, "year": 2025, "venue": "Journal of Computer Science and Technology Studies", "citations": 20, "notes": "This study sought to explore the transliteration of English nouns containing /g/ by Microsoft Copilot (MC) and Google Translate (GT), and find out which Arabic", "metrics": null }, { "id": "improving-mispronunciation-detection-and-diagnosis-for-non-n", "name": "Improving mispronunciation detection and diagnosis for non- native learners of the Arabic language", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "classification" ], "links": { "paper": "https://doi.org/10.1007/s10791-024-09489-8" }, "year": 2025, "venue": "Discover Computing", "citations": 20, "notes": "Mispronunciation detection and diagnosis (MDD) is a core component of computer-assisted pronunciation training (CAPT), which aims to provide opportunities for", "metrics": null }, { "id": "the-dialects-gap-a-multi-task-learning-approach-for-enhancin", "name": "The dialects gap: A multi-task learning approach for enhancing hate speech detection in Arabic dialects", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification" ], "links": { "paper": "https://doi.org/10.1016/j.eswa.2025.128584" }, "year": 2025, "venue": "Expert systems with applications", "citations": 20, "metrics": null }, { "id": "arabic-speech-recognition-using-neural-networks-concepts-lit", "name": "Arabic speech recognition using neural networks: concepts, literature review and challenges", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "survey", "asr" ], "links": { "paper": "https://doi.org/10.1007/s43994-025-00213-w" }, "year": 2025, "venue": "Journal of Umm Al-Qura University for Applied Sciences", "citations": 18, "notes": "The ability to recognize and translate human speech has grown in importance.", "metrics": null }, { "id": "radar-an-ensemble-approach-for-radicalization-detection-in-a", "name": "RADAR#: An Ensemble Approach for Radicalization Detection in Arabic Social Media Using Hybrid Deep Learning and Transformer Models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification" ], "links": { "paper": "https://doi.org/10.3390/info16070522" }, "year": 2025, "venue": "Inf.", "citations": 17, "notes": "RADAR#, a deep ensemble approach for the detection of radicalization in Arabic tweets, is introduced in this paper.", "metrics": null }, { "id": "translation-of-arabic-expressions-of-impossibility-by-ai-and", "name": "Translation of Arabic Expressions of Impossibility by AI and Student-Translators: A Comparative Study", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "evaluation" ], "links": { "paper": "https://doi.org/10.32996/jcsts.2025.7.8.33" }, "year": 2025, "venue": "Journal of Computer Science and Technology Studies", "citations": 17, "notes": "Expressions of impossibility (EIs) refer to events that never or rarely happen, things that are impossible to find, tasks that are difficult or impossible to", "metrics": null }, { "id": "a-survey-of-code-switched-arabic-nlp-progress-challenges-and", "name": "A Survey of Code-switched Arabic NLP: Progress, Challenges, and Future Directions", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey" ], "links": { "paper": "https://arxiv.org/abs/2501.13419" }, "year": 2025, "venue": "International Conference on Computational Linguistics", "citations": 16, "notes": "Language in the Arab world presents a complex diglossic and multilingual setting, involving the use of Modern Standard Arabic, various dialects and", "metrics": null }, { "id": "arabic-fake-news-dataset-development-humans-and-ai-generated", "name": "Arabic Fake News Dataset Development: Humans and AI-Generated Contributions", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification", "dataset" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2025.3556376" }, "year": 2025, "venue": "IEEE Access", "citations": 16, "notes": "The extensive use of social media platforms has promoted the rapid spread of fake news on the internet, such as fake reviews, rumors, and propaganda.", "metrics": null }, { "id": "evaluating-translation-quality-a-qualitative-and-quantitativ", "name": "Evaluating Translation Quality: A Qualitative and Quantitative Assessment of Machine and LLM-Driven Arabic-English Translations", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "pretraining", "evaluation" ], "links": { "paper": "https://doi.org/10.3390/info16060440" }, "year": 2025, "venue": "Inf.", "citations": 16, "notes": "This study investigates translation quality between Arabic and English, comparing traditional rule-based machine translation systems, modern neural machine", "metrics": null }, { "id": "a-comprehensive-survey-on-arabic-text-augmentation-approache", "name": "A comprehensive survey on Arabic text augmentation: approaches, challenges, and applications", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey" ], "links": { "paper": "https://doi.org/10.1007/s00521-025-11020-z" }, "year": 2025, "venue": "Neural computing & applications (Print)", "citations": 15, "notes": "Arabic is a linguistically complex language with a rich structure and valuable syntax that pose unique challenges for natural language processing (NLP)", "metrics": null }, { "id": "determining-the-meter-of-classical-arabic-poetry-using-deep", "name": "Determining the meter of classical Arabic poetry using deep learning: a performance analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://doi.org/10.3389/frai.2025.1523336" }, "year": 2025, "venue": "Frontiers Artif. Intell.", "citations": 15, "notes": "In this study, a deep learning-based approach was developed to accurately determine the meter of", "metrics": null }, { "id": "how-well-can-llms-grade-essays-in-arabic", "name": "How well can LLMs Grade Essays in Arabic?", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2501.16516" }, "year": 2025, "venue": "Computers and Education: Artificial Intelligence", "citations": 15, "notes": "This research assesses the effectiveness of state-of-the-art large language models (LLMs), including ChatGPT, Llama, Aya, Jais, and ACEGPT, in the task of", "metrics": null }, { "id": "leveraging-chatgpt-for-enhancing-arabic-nlp-application-for", "name": "Leveraging ChatGPT for Enhancing Arabic NLP: Application for Semantic Role Labeling and Cross-Lingual Annotation Projection", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2025.3525493" }, "year": 2025, "venue": "IEEE Access", "citations": 15, "notes": "Semantic role labeling involves assigning semantic roles to sentence arguments, providing rich information for various NLP tasks and applications.", "metrics": null }, { "id": "topic-modeling-and-sentiment-analysis-of-arabic-news-headlin", "name": "Topic Modeling and Sentiment Analysis of Arabic News Headlines for a Societal Well-Being Scoring and Monitoring System: Moroccan Use Case", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "classification" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2025.3538888" }, "year": 2025, "venue": "IEEE Access", "citations": 15, "notes": "In today’s information-rich society, the rapid dissemination of news content across digital platforms plays a pivotal role in shaping public perception and", "metrics": null }, { "id": "msa-at-arahealthqa-2025-shared-task-enhancing-llm-performance-for-arab", "name": "!MSA at AraHealthQA 2025 Shared Task: Enhancing LLM Performance for Arabic Clinical Question Answering through Prompt Engineering and Ensemble Learning", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa" ], "links": { "paper": "https://arxiv.org/abs/2509.11365" }, "dialects": [ "msa" ], "year": 2025, "venue": "arXiv 2025", "notes": "We present our systems for Track 2 (General Arabic Health QA, MedArabiQ) of the AraHealthQA-2025 shared task.", "metrics": null }, { "id": "msa-at-barec-shared-task-2025-ensembling-arabic-transformers-for-reada", "name": "!MSA at BAREC Shared Task 2025: Ensembling Arabic Transformers for Readability Assessment", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "nlp" ], "links": { "paper": "https://arxiv.org/abs/2509.10040" }, "dialects": [ "msa" ], "year": 2025, "venue": "arXiv 2025", "notes": "We present MSAs winning system for the BAREC 2025 Shared Task on fine-grained Arabic readability assessment, achieving first place in six of six tracks.", "metrics": null }, { "id": "3lm-bridging-arabic-stem-and-code-through-benchmarking", "name": "3LM: Bridging Arabic, STEM, and Code through Benchmarking", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "translation", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2507.15850" }, "dialects": [ "lev" ], "year": 2025, "venue": "arXiv 2025", "notes": "To help bridge this gap, we present 3LM, a suite of three benchmarks designed specifically for Arabic.", "metrics": null }, { "id": "a-culturally-diverse-multilingual-multimodal-video-benchmark-model", "name": "A Culturally-diverse Multilingual Multimodal Video Benchmark & Model", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "translation", "qa", "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2506.07032", "website": "https://mbzuai-oryx.github.io/ViMUL" }, "year": 2025, "venue": "arXiv 2025", "notes": "In pursuit of more inclusive video LMMs, we introduce a multilingual Video LMM benchmark, named ViMUL-Bench, to evaluate Video LMMs across 14 languages.", "metrics": null }, { "id": "a-large-and-balanced-corpus-for-fine-grained-arabic-readability-assess", "name": "A Large and Balanced Corpus for Fine-grained Arabic Readability Assessment", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2502.13520" }, "year": 2025, "venue": "arXiv 2025", "notes": "This paper introduces the Balanced Arabic Readability Evaluation Corpus (BAREC), a large-scale, fine-grained dataset for Arabic readability assessment.", "metrics": null }, { "id": "adapting-falcon3-7b-language-model-for-arabic-methods-challenges-and-o", "name": "Adapting Falcon3-7B Language Model for Arabic: Methods, Challenges, and Outcomes", "type": "paper", "country": "AE", "org": "TII (UAE)", "license": "unknown", "modality": "text", "tasks": [ "llm-adaptation" ], "links": { "paper": "https://aclanthology.org/2025.arabicnlp-main.1/" }, "year": 2025, "venue": "ArabicNLP 2025", "notes": "Describes adapting the Falcon3-7B language model to Arabic, from data collection through training and outcomes.", "metrics": null }, { "id": "adi-20-arabic-dialect-identification-dataset-and-models", "name": "ADI-20: Arabic Dialect Identification dataset and models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "asr", "dialect-id", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2511.10070" }, "dialects": [ "msa", "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "We present ADI-20, an extension of the previously published ADI-17 Arabic Dialect Identification (ADI) dataset.", "metrics": null }, { "id": "advancing-arabic-reverse-dictionary-systems-a-transformer-based-approa", "name": "Advancing Arabic Reverse Dictionary Systems: A Transformer-Based Approach with Dataset Construction Guidelines", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2504.21475" }, "year": 2025, "venue": "arXiv 2025", "notes": "We present a novel transformer-based approach with a semi-encoder neural network architecture featuring geometrically decreasing layers.", "metrics": null }, { "id": "ahasis-shared-task-on-sentiment-analysis-for-arabic-dialects", "name": "AHaSIS: Shared Task on Sentiment Analysis for Arabic Dialects", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "sentiment", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2511.13335" }, "dialects": [ "gulf", "magh", "msa", "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "The hospitality industry in the Arab world increasingly relies on customer feedback to shape services.", "metrics": null }, { "id": "alarb-an-arabic-legal-argument-reasoning-benchmark", "name": "ALARB: An Arabic Legal Argument Reasoning Benchmark", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "legal", "reasoning", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2510.00694" }, "year": 2025, "venue": "arXiv 2025", "notes": "ALARB: 13K Saudi commercial court cases for benchmarking Arabic LLM legal reasoning, with verdict prediction and regulation retrieval tasks.", "metrics": null }, { "id": "algerian-dialect", "name": "Algerian Dialect", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://arxiv.org/abs/2512.19543", "website": "https://doi.org/10.17632/zzwg3nnhsz.2" }, "dialects": [ "magh" ], "year": 2025, "venue": "arXiv 2025", "notes": "We present Algerian Dialect, a large-scale sentiment-annotated dataset consisting of 45,000 YouTube comments written in Algerian Arabic dialect.", "metrics": null }, { "id": "alhd-a-large-scale-and-multigenre-benchmark-dataset-for-arabic-llm-gen", "name": "ALHD: A Large-Scale and Multigenre Benchmark Dataset for Arabic LLM-Generated Text Detection", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2510.03502" }, "dialects": [ "msa" ], "year": 2025, "venue": "arXiv 2025", "notes": "We introduce ALHD, the first large-scale comprehensive Arabic dataset explicitly designed to distinguish between human- and LLM-generated texts.", "metrics": null }, { "id": "alignar-generative-sentence-alignment-for-arabic-english-parallel-corp", "name": "AlignAR: Generative Sentence Alignment for Arabic-English Parallel Corpora of Legal and Literary Texts", "type": "paper", "country": "INTL", "org": "xxx", "license": "unknown", "modality": "text", "tasks": [ "translation", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2512.21842", "github": "https://github.com/XXX" }, "year": 2025, "venue": "arXiv 2025", "notes": "We present AlignAR, a generative sentence alignment method, and a new Arabic-English dataset comprising simple legal and complex literary parallel texts.", "metrics": null }, { "id": "amcrawl-an-arabic-web-scale-dataset-of-interleaved-image-text-document", "name": "AMCrawl: An Arabic Web-Scale Dataset of Interleaved Image-Text Documents and Image-Text Pairs", "type": "paper", "country": "SA", "org": "SDAIA", "license": "unknown", "modality": "multimodal", "tasks": [ "multimodal", "pretraining" ], "links": { "paper": "https://aclanthology.org/2025.arabicnlp-main.37/" }, "year": 2025, "venue": "ArabicNLP 2025", "notes": "First native Arabic multimodal web dataset of interleaved image-text documents and image-text pairs, filtered from Common Crawl for quality and safety.", "metrics": null }, { "id": "an-annotated-corpus-of-arabic-tweets-for-hate-speech-analysis", "name": "An Annotated Corpus of Arabic Tweets for Hate Speech Analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2505.11969" }, "year": 2025, "venue": "arXiv 2025", "notes": "Identifying hate speech content in the Arabic language is challenging due to the rich quality of dialectal variations.", "metrics": null }, { "id": "an-exploration-of-knowledge-editing-for-arabic", "name": "An Exploration of Knowledge Editing for Arabic", "type": "paper", "country": "QA", "org": "QCRI", "license": "unknown", "modality": "text", "tasks": [ "morphology" ], "links": { "paper": "https://aclanthology.org/2025.arabicnlp-main.34/" }, "year": 2025, "venue": "ArabicNLP 2025", "notes": "We present the first study of Arabic KE.", "metrics": null }, { "id": "an-open-research-dataset-of-the-1932-cairo-congress-of-arab-music", "name": "An Open Research Dataset of the 1932 Cairo Congress of Arab Music", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "music", "audio", "dataset" ], "links": { "paper": "https://arxiv.org/abs/2506.14503" }, "year": 2025, "venue": "arXiv 2025", "notes": "ORD-CC32: open dataset from the 1932 Cairo Congress of Arab Music recordings with maqam and iqa tags, tonic labels and acoustic features.", "metrics": null }, { "id": "ara-hope-human-centric-post-editing-evaluation-for-dialectal-arabic-to", "name": "Ara-HOPE: Human-Centric Post-Editing Evaluation for Dialectal Arabic to Modern Standard Arabic Translation", "type": "paper", "country": "INTL", "org": "abdullahalabdullah", "license": "unknown", "modality": "text", "tasks": [ "translation", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2512.21787", "github": "https://github.com/abdullahalabdullah/Ara-HOPE" }, "dialects": [ "msa", "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "This paper introduces Ara-HOPE, a human-centric post-editing evaluation framework designed to systematically address these challenges.", "metrics": null }, { "id": "arabemonet-a-lightweight-hybrid-2d-cnn-bilstm-model-with-attention-for", "name": "ArabEmoNet: A Lightweight Hybrid 2D CNN-BiLSTM Model with Attention for Robust Arabic Speech Emotion Recognition", "type": "paper", "country": "AE", "org": "MBZUAI", "license": "unknown", "modality": "speech", "tasks": [ "emotion" ], "links": { "paper": "https://aclanthology.org/2025.arabicnlp-main.17/" }, "year": 2025, "venue": "ArabicNLP 2025", "notes": "We introduce ArabEmoNet, a lightweight architecture designed to overcome these limitations and deliver state-of-the-art performance.", "metrics": null }, { "id": "arabic-asr-on-the-sada-large-scale-arabic-speech-corpus-with-transform", "name": "Arabic ASR on the SADA Large-Scale Arabic Speech Corpus with Transformer-Based Models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "evaluation", "speech", "vision" ], "links": { "paper": "https://arxiv.org/abs/2508.12968" }, "dialects": [ "gulf", "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "We explore the performance of several state-of-the-art automatic speech recognition (ASR) models on a large-scale Arabic speech dataset, the SADA.", "metrics": null }, { "id": "arabic-little-stt-arabic-children-speech-recognition-dataset", "name": "Arabic Little STT: Arabic Children Speech Recognition Dataset", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2510.23319" }, "dialects": [ "lev" ], "year": 2025, "venue": "arXiv 2025", "notes": "We present our created dataset, Arabic Little STT, a dataset of Levantine Arabic child speech recorded in classrooms.", "metrics": null }, { "id": "arabic-multimodal-machine-learning-datasets-applications-approaches-an", "name": "Arabic Multimodal Machine Learning: Datasets, Applications, Approaches, and Challenges", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "embedding", "sentiment", "survey", "speech", "vision" ], "links": { "paper": "https://arxiv.org/abs/2508.12227" }, "year": 2025, "venue": "arXiv 2025", "notes": "Multimodal Machine Learning (MML) aims to integrate and analyze information from diverse modalities, such as text, audio, and visuals.", "metrics": null }, { "id": "arabic-tool-calling-data-fanar", "name": "Arabic tool-calling data (Fanar)", "type": "paper", "country": "QA", "org": "QCRI", "license": "unknown", "modality": "text", "tasks": [ "tool-calling" ], "links": { "paper": "https://arxiv.org/abs/2509.20957" }, "year": 2025, "notes": "Paper on data strategies and instruction tuning for tool calling in Arabic LLMs (Fanar).", "metrics": null }, { "id": "arabjobs-a-multinational-corpus-of-arabic-job-ads", "name": "ArabJobs: A Multinational Corpus of Arabic Job Ads", "type": "paper", "country": "INTL", "org": "drelhaj", "license": "unknown", "modality": "text", "tasks": [ "jobs", "bias", "classification" ], "links": { "paper": "https://arxiv.org/abs/2509.22589", "github": "https://github.com/drelhaj/ArabJobs" }, "year": 2025, "venue": "arXiv 2025", "notes": "ArabJobs: 8,500 Arabic job ads from Egypt, Jordan, Saudi Arabia and the UAE for gender-bias, profession and salary analysis.", "metrics": null }, { "id": "arafinnews-arabic-financial-summarisation-with-domain-adapted-llms", "name": "AraFinNews: Arabic Financial Summarisation with Domain-Adapted LLMs", "type": "paper", "country": "INTL", "org": "arabicnlp-uk", "license": "unknown", "modality": "text", "tasks": [ "summarization", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2511.01265", "github": "https://github.com/ArabicNLP-uk/AraFinNews" }, "year": 2025, "venue": "arXiv 2025", "notes": "We introduce AraFinNews, the largest publicly available Arabic financial news dataset to date.", "metrics": null }, { "id": "arahallueval-a-fine-grained-hallucination-evaluation-framework-for-ara", "name": "AraHalluEval: A Fine-grained Hallucination Evaluation Framework for Arabic LLMs", "type": "paper", "country": "INTL", "org": "aishaalansari57", "license": "unknown", "modality": "text", "tasks": [ "qa", "summarization", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2509.04656", "github": "https://github.com/aishaalansari57/AraHalluEval" }, "year": 2025, "venue": "arXiv 2025", "notes": "This paper presents the first comprehensive hallucination evaluation of Arabic and multilingual LLMs.", "metrics": null }, { "id": "arahealthqa-2025-the-first-shared-task-on-arabic-health-question-answe", "name": "AraHealthQA 2025: The First Shared Task on Arabic Health Question Answering", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa", "summarization", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2508.20047" }, "year": 2025, "venue": "arXiv 2025", "notes": "We introduce AraHealthQA 2025, the Comprehensive Arabic Health Question Answering Shared Task, held in conjunction with ArabicNLP 2025.", "metrics": null }, { "id": "aralingbench-a-human-annotated-benchmark-for-evaluating-arabic-linguis", "name": "AraLingBench A Human-Annotated Benchmark for Evaluating Arabic Linguistic Capabilities of Large Language Models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2511.14295" }, "year": 2025, "venue": "arXiv 2025", "notes": "We present AraLingBench: a fully human annotated benchmark for evaluating the Arabic linguistic competence of large language models (LLMs).", "metrics": null }, { "id": "arareasoner-evaluating-reasoning-based-llms-for-arabic-nlp", "name": "AraReasoner: Evaluating Reasoning-Based LLMs for Arabic NLP", "type": "paper", "country": "INTL", "org": "gufransabri", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2506.08768", "github": "https://github.com/gufranSabri/deepseek-evals" }, "dialects": [ "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "This paper presents a comprehensive benchmarking study of multiple reasoning-focused LLMs.", "metrics": null }, { "id": "aratable-benchmarking-llms-reasoning-and-understanding-of-arabic-tabul", "name": "AraTable: Benchmarking LLMs' Reasoning and Understanding of Arabic Tabular Data", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2507.18442" }, "year": 2025, "venue": "arXiv 2025", "notes": "We present AraTable, a novel and comprehensive benchmark designed to evaluate the reasoning and understanding capabilities of LLMs when applied.", "metrics": null }, { "id": "arb-a-comprehensive-arabic-multimodal-reasoning-benchmark", "name": "ARB: A Comprehensive Arabic Multimodal Reasoning Benchmark", "type": "paper", "country": "AE", "org": "MBZUAI", "license": "unknown", "modality": "multimodal", "tasks": [ "ocr", "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2505.17021", "github": "https://github.com/mbzuai-oryx/ARB" }, "year": 2025, "venue": "arXiv 2025", "notes": "We introduce the Comprehensive Arabic Multimodal Reasoning Benchmark.", "metrics": null }, { "id": "arfake-a-robust-framework-for-multi-dialect-arabic-speech-spoofing-det", "name": "ArFake: A Robust Framework for Multi-Dialect Arabic Speech Spoofing Detection Benchmark", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "spoof-detection", "speech", "tts" ], "links": { "paper": "https://arxiv.org/abs/2509.22808" }, "dialects": [ "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "ArFake: first multi-dialect Arabic spoofed-speech dataset built from several TTS models, with a spoof-detection evaluation pipeline.", "metrics": null }, { "id": "arvoice-a-multi-speaker-dataset-for-arabic-speech-synthesis", "name": "ArVoice: A Multi-Speaker Dataset for Arabic Speech Synthesis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "tts", "diacritization", "speech" ], "links": { "paper": "https://arxiv.org/abs/2505.20506" }, "dialects": [ "msa" ], "year": 2025, "venue": "arXiv 2025", "notes": "We introduce ArVoice, a multi-speaker Modern Standard Arabic.", "metrics": null }, { "id": "arzen-multigenre-an-aligned-parallel-dataset-of-egyptian-arabic-song-l", "name": "ArzEn-MultiGenre: An aligned parallel dataset of Egyptian Arabic song lyrics, novels, and subtitles, with English translations", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2508.01411" }, "dialects": [ "egy" ], "year": 2025, "venue": "arXiv 2025", "notes": "ArzEn-MultiGenre is a parallel dataset of Egyptian Arabic song lyrics, novels.", "metrics": null }, { "id": "autoarabic-a-three-stage-framework-for-localizing-video-text-retrieval", "name": "AutoArabic: A Three-Stage Framework for Localizing Video-Text Retrieval Benchmarks", "type": "paper", "country": "INTL", "org": "tahaalshatiri", "license": "unknown", "modality": "text", "tasks": [ "embedding", "translation", "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2509.16438", "github": "https://github.com/Tahaalshatiri/AutoArabic" }, "dialects": [ "msa" ], "year": 2025, "venue": "arXiv 2025", "notes": "We introduce a three-stage framework, AutoArabic, utilizing state-of-the-art large language models.", "metrics": null }, { "id": "automatic-pronunciation-error-detection-and-correction-of-the-holy-qur", "name": "Automatic Pronunciation Error Detection and Correction of the Holy Quran's Learners Using Deep Learning", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "diacritization", "evaluation", "speech", "quran" ], "links": { "paper": "https://arxiv.org/abs/2509.00094", "website": "https://obadx.github.io/quran-muaalem/en" }, "dialects": [ "classical", "msa" ], "year": 2025, "venue": "arXiv 2025", "notes": "We release our work as open-source: https://obadx.github.io/quran-muaalem/en/", "metrics": null }, { "id": "balsam-a-platform-for-benchmarking-arabic-large-language-models", "name": "BALSAM: A Platform for Benchmarking Arabic Large Language Models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2507.22603" }, "dialects": [ "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "In particular, we introduce BALSAM, a comprehensive, community-driven benchmark aimed at advancing Arabic LLM development and evaluation.", "metrics": null }, { "id": "benchmarking-the-legal-reasoning-of-llms-in-arabic-islamic-inheritance", "name": "Benchmarking the Legal Reasoning of LLMs in Arabic Islamic Inheritance Cases", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2508.15796" }, "year": 2025, "venue": "arXiv 2025", "notes": "Islamic inheritance domain holds significant importance for Muslims to ensure fair distribution of shares between heirs.", "metrics": null }, { "id": "benchmarking-the-medical-understanding-and-reasoning-of-large-language", "name": "Benchmarking the Medical Understanding and Reasoning of Large Language Models in Arabic Healthcare Tasks", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2508.15797" }, "year": 2025, "venue": "arXiv 2025", "notes": "Recent progress in large language models (LLMs) has showcased impressive proficiency in numerous Arabic natural language processing (NLP) applications.", "metrics": null }, { "id": "beyond-mcq-an-open-ended-arabic-cultural-qa-benchmark-with-dialect-var", "name": "Beyond MCQ: An Open-Ended Arabic Cultural QA Benchmark with Dialect Variants", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "qa", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2510.24328" }, "dialects": [ "msa", "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "We propose a comprehensive method that (i) translates Modern Standard Arabic (MSA) multiple-choice questions.", "metrics": null }, { "id": "building-and-aligning-comparable-corpora", "name": "Building and Aligning Comparable Corpora", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2508.02555" }, "year": 2025, "venue": "arXiv 2025", "notes": "In this paper, we present a method to build comparable corpora from Wikipedia encyclopedia and EURONEWS website in English, French and Arabic languages.", "metrics": null }, { "id": "busted-at-arageneval-shared-task-a-comparative-study-of-transformer-ba", "name": "BUSTED at AraGenEval Shared Task: A Comparative Study of Transformer-Based Models for Arabic AI-Generated Text Detection", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2510.20610" }, "year": 2025, "venue": "arXiv 2025", "notes": "This paper details our submission to the AraGenEval Shared Task on Arabic AI-generated text detection, where our team, BUSTED, secured 5th place.", "metrics": null }, { "id": "capturing-intra-dialectal-variation-in-qatari-arabic-a-corpus-of-cultu", "name": "Capturing Intra-Dialectal Variation in Qatari Arabic: A Corpus of Cultural and Gender Dimensions", "type": "paper", "country": "QA", "org": "Hamad Bin Khalifa University", "license": "unknown", "modality": "text", "tasks": [ "asr", "pronunciation" ], "links": { "paper": "https://aclanthology.org/2025.arabicnlp-main.18/" }, "dialects": [ "gulf" ], "year": 2025, "venue": "ArabicNLP 2025", "notes": "We present the first publicly available, multidimensional corpus of Qatari Arabic that captures intra-dialectal variation across Urban and Bedouin speakers.", "metrics": null }, { "id": "carma-comprehensive-automatically-annotated-reddit-mental-health-datas", "name": "CARMA: Comprehensive Automatically-annotated Reddit Mental Health Dataset for Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "nlp" ], "links": { "paper": "https://arxiv.org/abs/2511.03102" }, "year": 2025, "venue": "arXiv 2025", "notes": "We present CARMA, the first automatically annotated large-scale dataset of Arabic Reddit posts.", "metrics": null }, { "id": "comparative-approaches-to-sentiment-analysis-using-datasets-in-major-e", "name": "Comparative Approaches to Sentiment Analysis Using Datasets in Major European and Arabic Languages", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://arxiv.org/abs/2501.12540" }, "year": 2025, "venue": "arXiv 2025", "notes": "This study explores transformer-based models such as BERT, mBERT, and XLM-R for multi-lingual sentiment analysis across diverse linguistic structures.", "metrics": null }, { "id": "cross-lingual-synthdocs-a-large-scale-synthetic-corpus-for-any-to-arab", "name": "Cross-Lingual SynthDocs: A Large-Scale Synthetic Corpus for Any to Arabic OCR and Document Understanding", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "ocr", "diacritization", "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2511.04699" }, "year": 2025, "venue": "arXiv 2025", "notes": "Cross-Lingual SynthDocs is a large-scale synthetic corpus designed to address the scarcity of Arabic resources for Optical Character Recognition.", "metrics": null }, { "id": "cs-fleurs-a-massively-multilingual-and-code-switched-speech-dataset", "name": "CS-FLEURS: A Massively Multilingual and Code-Switched Speech Dataset", "type": "paper", "country": "INTL", "org": "byan", "license": "cc-by-nc-4.0", "modality": "speech", "tasks": [ "asr", "tts", "translation", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2509.14161", "hf": "https://huggingface.co/datasets/byan/cs-fleurs" }, "year": 2025, "venue": "arXiv 2025", "notes": "We present CS-FLEURS, a new dataset for developing and evaluating code-switched speech recognition and translation systems beyond high-resourced languages.", "metrics": { "downloads": 802, "likes": 19, "lastModified": "2025-09-17" } }, { "id": "cvpd-at-qias-2025-shared-task-an-efficient-encoder-based-approach-for", "name": "CVPD at QIAS 2025 Shared Task: An Efficient Encoder-Based Approach for Islamic Inheritance Reasoning", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2509.00457" }, "year": 2025, "venue": "arXiv 2025", "notes": "We present a lightweight framework for solving multiple-choice inheritance questions using a specialised Arabic text encoder and Attentive Relevance Scoring.", "metrics": null }, { "id": "dialect2sql-a-novel-text-to-sql-dataset-for-arabic-dialects-with-a-foc", "name": "Dialect2SQL: A Novel Text-to-SQL Dataset for Arabic Dialects with a Focus on Moroccan Darija", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2501.11498" }, "dialects": [ "magh", "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "In this work, we introduce Dialect2SQL, the first large-scale, cross-domain text-to-SQL dataset in an Arabic dialect.", "metrics": null }, { "id": "dialectalarabicmmlu-benchmarking-dialectal-capabilities-in-arabic-and", "name": "DialectalArabicMMLU: Benchmarking Dialectal Capabilities in Arabic and Multilingual Language Models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "qa", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2510.27543" }, "dialects": [ "egy", "gulf", "lev", "magh", "msa", "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "We present DialectalArabicMMLU, a new benchmark for evaluating the performance of large language models (LLMs) across Arabic dialects.", "metrics": null }, { "id": "dialg2p-dialectal-grapheme-to-phoneme-arabic-as-a-case-study", "name": "DialG2P: Dialectal Grapheme-to-Phoneme. Arabic as a Case Study", "type": "paper", "country": "QA", "org": "QCRI", "license": "unknown", "modality": "text", "tasks": [ "g2p" ], "links": { "paper": "https://aclanthology.org/2025.arabicnlp-main.38/", "github": "https://github.com/qcri/DialG2P" }, "dialects": [ "egy" ], "year": 2025, "venue": "ArabicNLP 2025", "notes": "We introduce an end-to-end dialectal G2P for Egyptian Arabic, a dialect without standard orthography.", "metrics": null }, { "id": "emohopespeech-an-annotated-dataset-of-emotions-and-hope-speech-in-engl", "name": "EmoHopeSpeech: An Annotated Dataset of Emotions and Hope Speech in English and Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2505.11959" }, "year": 2025, "venue": "arXiv 2025", "notes": "This research introduces a bilingual dataset comprising 23,456 entries for Arabic and 10,036 entries for English.", "metrics": null }, { "id": "enhanced-arabic-text-retrieval-with-attentive-relevance-scoring", "name": "Enhanced Arabic Text Retrieval with Attentive Relevance Scoring", "type": "paper", "country": "INTL", "org": "bekhouche", "license": "unknown", "modality": "text", "tasks": [ "embedding", "diacritization", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2507.23404", "github": "https://github.com/Bekhouche/APR" }, "dialects": [ "msa", "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "In this paper, we present an enhanced Dense Passage Retrieval (DPR) framework developed specifically for Arabic.", "metrics": null }, { "id": "evaluating-arabic-llms-a-survey-of-benchmarks-methods-and", "name": "Evaluating Arabic LLMs: A Survey of Benchmarks, Methods, and Gaps", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "year": 2025, "venue": "arXiv 2025", "tasks": [ "survey", "benchmark", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2510.13430" }, "notes": "Survey of Arabic LLM benchmarks, evaluation methods and gaps.", "metrics": null }, { "id": "evaluating-prompt-relevance-in-arabic-automatic-essay-scoring-insights", "name": "Evaluating Prompt Relevance in Arabic Automatic Essay Scoring: Insights from Synthetic and Real-World Data", "type": "paper", "country": "AE", "org": "NYU Abu Dhabi", "license": "unknown", "modality": "text", "tasks": [ "essay-scoring" ], "links": { "paper": "https://aclanthology.org/2025.arabicnlp-main.13/" }, "year": 2025, "venue": "ArabicNLP 2025", "notes": "We present the first systematic study of binary prompt-essay relevance classification, supporting both AES scoring and dataset annotation.", "metrics": null }, { "id": "fann-or-flop-a-multigenre-multiera-benchmark-for-arabic-poetry-underst", "name": "Fann or Flop: A Multigenre, Multiera Benchmark for Arabic Poetry Understanding in LLMs", "type": "paper", "country": "AE", "org": "MBZUAI", "license": "unknown", "modality": "text", "tasks": [ "evaluation", "poetry" ], "links": { "paper": "https://arxiv.org/abs/2505.18152", "github": "https://github.com/mbzuai-oryx/FannOrFlop" }, "dialects": [ "classical" ], "year": 2025, "venue": "arXiv 2025", "notes": "Arabic poetry is one of the richest and most culturally rooted forms of expression in the Arabic language, known for its layered meanings.", "metrics": null }, { "id": "feature-engineering-is-not-dead-a-step-towards-state-of-the-art-for-ar", "name": "Feature Engineering is not Dead: A Step Towards State of the Art for Arabic Automated Essay Scoring", "type": "paper", "country": "QA", "org": "Qatar University", "license": "unknown", "modality": "text", "tasks": [ "essay-scoring" ], "links": { "paper": "https://aclanthology.org/2025.arabicnlp-main.19/", "github": "https://github.com/Maroibo/AES_features" }, "year": 2025, "venue": "ArabicNLP 2025", "notes": "Experiments are conducted on a dataset of 620 Arabic essays, each annotated with both holistic and trait-specific scores.", "metrics": null }, { "id": "gate-general-arabic-text-embedding-for-enhanced-sts", "name": "GATE: General Arabic Text Embedding for Enhanced STS", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "year": 2025, "venue": "arXiv 2025", "tasks": [ "embedding", "sts" ], "links": { "paper": "https://arxiv.org/abs/2505.24581" }, "notes": "Matryoshka-style general Arabic text embeddings for semantic textual similarity.", "metrics": null }, { "id": "hybrid-deep-learning-and-signal-processing-for-arabic-dialect-recognit", "name": "Hybrid Deep Learning and Signal Processing for Arabic Dialect Recognition in Low-Resource Settings", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2506.21386" }, "dialects": [ "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "Arabic dialect recognition presents a significant challenge in speech technology due to the linguistic diversity of Arabic and the scarcity.", "metrics": null }, { "id": "iqra-eval-a-shared-task-on-qur-anic-pronunciation-assessment", "name": "Iqra’Eval: A Shared Task on Qur’anic Pronunciation Assessment", "type": "paper", "country": "INTL", "org": "University of Sheffield", "license": "unknown", "modality": "text", "tasks": [ "pronunciation", "shared-task" ], "links": { "paper": "https://aclanthology.org/2025.arabicnlp-sharedtasks.61/", "hf": "https://huggingface.co/spaces/IqraEval/" }, "dialects": [ "msa" ], "year": 2025, "venue": "ArabicNLP 2025", "notes": "We present the findings of the first shared task on Qur’anic pronunciation assessment, which focuses on addressing the unique challenges of evaluating.", "metrics": null }, { "id": "jawaher-a-multidialectal-dataset-of-arabic-proverbs-for-llm-benchmarki", "name": "Jawaher: A Multidialectal Dataset of Arabic Proverbs for LLM Benchmarking", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2503.00231" }, "dialects": [ "lev", "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "To address this, we introduce Jawaher, a benchmark designed to assess LLMs' capacity to comprehend and interpret Arabic proverbs.", "metrics": null }, { "id": "kitab-bench-a-comprehensive-multi-domain-benchmark-for-arabic-ocr-and", "name": "KITAB-Bench: A Comprehensive Multi-Domain Benchmark for Arabic OCR and Document Understanding", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "ocr", "embedding", "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2502.14949" }, "year": 2025, "venue": "arXiv 2025", "notes": "We present KITAB-Bench, a comprehensive Arabic OCR benchmark that fills the gaps in current evaluation systems.", "metrics": null }, { "id": "konooz-multi-domain-multi-dialect-corpus-for-named-entity-recognition", "name": "Konooz: Multi-domain Multi-dialect Corpus for Named Entity Recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "ner", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2506.12615", "website": "https://sina.birzeit.edu/wojood/#download" }, "dialects": [ "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "We introduce Konooz, a novel multi-dimensional corpus covering 16 Arabic dialects across 10 domains, resulting in 160 distinct corpora.", "metrics": null }, { "id": "laila-a-large-trait-based-dataset-for-arabic-automated-essay-scoring", "name": "LAILA: A Large Trait-Based Dataset for Arabic Automated Essay Scoring", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2512.24235" }, "year": 2025, "venue": "arXiv 2025", "notes": "We introduce LAILA, the largest publicly available Arabic AES dataset to date.", "metrics": null }, { "id": "linto-audio-and-textual-datasets-to-train-and-evaluate-automatic-speec", "name": "LinTO Audio and Textual Datasets to Train and Evaluate Automatic Speech Recognition in Tunisian Arabic Dialect", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2504.02604" }, "dialects": [ "magh" ], "year": 2025, "venue": "arXiv 2025", "notes": "We propose the LinTO audio and textual datasets -- comprehensive resources that capture phonological and lexical features of Tunisian Arabic Dialect.", "metrics": null }, { "id": "llmvox-autoregressive-streaming-text-to-speech-model-for-any-llm", "name": "LLMVoX: Autoregressive Streaming Text-to-Speech Model for Any LLM", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "chat", "tts", "speech", "vision" ], "links": { "paper": "https://arxiv.org/abs/2503.04724", "website": "https://mbzuai-oryx.github.io/LLMVoX" }, "year": 2025, "venue": "arXiv 2025", "notes": "In contrast, we propose LLMVoX, a lightweight 30M-parameter, LLM-agnostic, autoregressive streaming TTS system that generates high-quality speech.", "metrics": null }, { "id": "macosworld-a-multilingual-interactive-benchmark-for-gui-agents", "name": "macOSWorld: A Multilingual Interactive Benchmark for GUI Agents", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2506.04135", "website": "https://macos-world.github.io" }, "year": 2025, "venue": "arXiv 2025", "notes": "To bridge the gaps, we present macOSWorld, the first comprehensive benchmark for evaluating GUI agents.", "metrics": null }, { "id": "maproc-at-ahasis-shared-task-few-shot-and-sentence-transformer-for-sen", "name": "MAPROC at AHaSIS Shared Task: Few-Shot and Sentence Transformer for Sentiment Analysis of Arabic Hotel Reviews", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2511.15291" }, "dialects": [ "gulf", "magh", "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "This paper describes our approach to the AHaSIS shared task, which focuses on sentiment analysis on Arabic dialects in the hospitality domain.", "metrics": null }, { "id": "masrad-arabic-terminology-management-corpora-with-semi-automatic-const", "name": "MASRAD: Arabic Terminology Management Corpora with Semi-Automatic Construction", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation" ], "links": { "paper": "https://arxiv.org/abs/2503.19211" }, "year": 2025, "venue": "arXiv 2025", "notes": "This paper presents MASRAD, a terminology dataset for Arabic terminology management, and a method with supporting tools for its semi-automatic construction.", "metrics": null }, { "id": "medarabiq-benchmarking-large-language-models-on-arabic-medical-tasks", "name": "MedArabiQ: Benchmarking Large Language Models on Arabic Medical Tasks", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2505.03427" }, "year": 2025, "venue": "arXiv 2025", "notes": "Large Language Models (LLMs) have demonstrated significant promise for various applications in healthcare.", "metrics": null }, { "id": "memeintel-explainable-detection-of-propagandistic-and-hateful-memes", "name": "MemeIntel: Explainable Detection of Propagandistic and Hateful Memes", "type": "paper", "country": "INTL", "org": "mohamedbayan", "license": "unknown", "modality": "multimodal", "tasks": [ "vision" ], "links": { "paper": "https://arxiv.org/abs/2502.16612", "github": "https://github.com/MohamedBayan/MemeIntel" }, "year": 2025, "venue": "arXiv 2025", "notes": "We introduce MemeXplain, an explanation-enhanced dataset for propagandistic memes in Arabic and hateful memes in English.", "metrics": null }, { "id": "mind-the-gap-a-review-of-arabic-post-training-datasets-and-their-limit", "name": "Mind the Gap: A Review of Arabic Post-Training Datasets and Their Limitations", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "qa", "summarization", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2507.14688" }, "year": 2025, "venue": "arXiv 2025", "notes": "This paper presents a review of publicly available Arabic post-training datasets on the Hugging Face Hub, organized along four key dimensions.", "metrics": null }, { "id": "mix-minhash-and-match-cross-source-agreement-for-multilingual-pretrain", "name": "Mix, MinHash, and Match: Cross-Source Agreement for Multilingual Pretraining Datasets", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2512.18834", "hf": "https://huggingface.co/collections/AdaMLLab/mixminmatch" }, "year": 2025, "venue": "arXiv 2025", "notes": "In this work, we propose to use this wasteful redundancy as a quality signal to create high-quality pretraining datasets.", "metrics": null }, { "id": "mizanqa-benchmarking-large-language-models-on-moroccan-legal-question", "name": "MizanQA: Benchmarking Large Language Models on Moroccan Legal Question Answering", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2508.16357" }, "dialects": [ "magh", "msa" ], "year": 2025, "venue": "arXiv 2025", "notes": "This paper introduces MizanQA (pronounced Mizan, meaning \"scale\" in Arabic, a universal symbol of justice).", "metrics": null }, { "id": "mole-metadata-extraction-and-validation-in-scientific-papers-using-llm", "name": "MOLE: Metadata Extraction and Validation in Scientific Papers Using LLMs", "type": "paper", "country": "SA", "org": "IVUL-KAUST", "license": "apache-2.0", "modality": "text", "tasks": [ "metadata-extraction", "llm" ], "links": { "paper": "https://arxiv.org/abs/2505.19800", "github": "https://github.com/IVUL-KAUST/MOLE", "hf": "https://huggingface.co/datasets/IVUL-KAUST/MOLE" }, "year": 2025, "venue": "arXiv 2025", "notes": "MOLE: LLM framework that extracts and validates metadata attributes from papers on non-Arabic language datasets, with a new benchmark.", "metrics": { "downloads": 49, "likes": 1, "lastModified": "2025-09-20" } }, { "id": "morphbpe-a-morpho-aware-tokenizer-bridging-linguistic-complexity-for-e", "name": "MorphBPE: A Morpho-Aware Tokenizer Bridging Linguistic Complexity for Efficient LLM Training Across Morphologies", "type": "paper", "country": "INTL", "org": "llm-lab-org", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2502.00894", "github": "https://github.com/llm-lab-org/MorphBPE" }, "year": 2025, "venue": "arXiv 2025", "notes": "We introduce MorphBPE, a morphology-aware extension of BPE that integrates linguistic structure into subword tokenization.", "metrics": null }, { "id": "mudric-multi-dialect-reasoning-for-arabic-commonsense-validation", "name": "MuDRiC: Multi-Dialect Reasoning for Arabic Commonsense Validation", "type": "paper", "country": "INTL", "org": "kareemelozeiri", "license": "unknown", "modality": "speech", "tasks": [ "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2508.13130", "github": "https://github.com/KareemElozeiri/MuDRiC" }, "dialects": [ "msa", "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "We introduce MuDRiC, an extended Arabic commonsense dataset incorporating multiple dialects.", "metrics": null }, { "id": "multi-agent-interactive-question-generation-framework-for-long-documen", "name": "Multi-Agent Interactive Question Generation Framework for Long Document Understanding", "type": "paper", "country": "INTL", "org": "wangk0b", "license": "unknown", "modality": "multimodal", "tasks": [ "qa", "vision" ], "links": { "paper": "https://arxiv.org/abs/2507.20145" }, "year": 2025, "venue": "arXiv 2025", "notes": "We propose a fully automated, multi-agent interactive framework to generate long-context questions efficiently.", "metrics": null }, { "id": "multi-lingual-cyber-threat-detection-in-tweets-x-using-ml-dl-and-llm-a", "name": "Multi-Lingual Cyber Threat Detection in Tweets/X Using ML, DL, and LLM: A Comparative Analysis", "type": "paper", "country": "INTL", "org": "mmurrad", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2502.04346", "github": "https://github.com/Mmurrad/Tweet-Data-Classification.git" }, "year": 2025, "venue": "arXiv 2025", "notes": "Cyber threat detection has become an important area of focus in today's digital age due to the growing spread of fake information and harmful content.", "metrics": null }, { "id": "multiprose-a-multi-label-arabic-dataset-for-propaganda-sentiment-and-e", "name": "MultiProSE: A Multi-label Arabic Dataset for Propaganda, Sentiment, and Emotion Detection", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2502.08319" }, "year": 2025, "venue": "arXiv 2025", "notes": "Propaganda is a form of persuasion that has been used throughout history with the intention goal of influencing people's opinions through rhetorical.", "metrics": null }, { "id": "munsit-at-nadi-2025-shared-task-2-pushing-the-boundaries-of-multidiale", "name": "Munsit at NADI 2025 Shared Task 2: Pushing the Boundaries of Multidialectal Arabic ASR with Weakly Supervised Pretraining and Continual Supervised Fine-tuning", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "dialect-id", "pretraining", "speech", "vision" ], "links": { "paper": "https://arxiv.org/abs/2508.08912" }, "dialects": [ "msa", "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "In this work, we present a scalable training pipeline that combines weakly supervised learning with supervised fine-tuning to develop a robust Arabic ASR model.", "metrics": null }, { "id": "nilechat-towards-linguistically-diverse-and-culturally-aware-llms-for", "name": "NileChat: Towards Linguistically Diverse and Culturally Aware LLMs for Local Communities", "type": "paper", "country": "INTL", "org": "UBC-NLP", "license": "unknown", "modality": "text", "tasks": [ "embedding", "translation", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2505.18383", "github": "https://github.com/UBC-NLP/nilechat" }, "dialects": [ "egy", "magh", "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "This work proposes a methodology to create both synthetic and retrieval-based pre-training data tailored to a specific community, considering.", "metrics": null }, { "id": "oasis-a-multilingual-and-multimodal-dataset-for-culturally-grounded-sp", "name": "OASIS: A Multilingual and Multimodal Dataset for Culturally Grounded Spoken Visual QA", "type": "paper", "country": "QA", "org": "QCRI", "license": "cc-by-nc-sa-4.0", "modality": "multimodal", "tasks": [ "qa", "evaluation", "speech", "vision" ], "links": { "paper": "https://arxiv.org/abs/2510.06371", "hf": "https://huggingface.co/datasets/QCRI/OASIS" }, "dialects": [ "msa" ], "year": 2025, "venue": "arXiv 2025", "notes": "We introduce OASIS, a large-scale culturally grounded multimodal QA dataset covering images, text, and speech.", "metrics": { "downloads": 50, "likes": 0, "lastModified": "2026-05-16" } }, { "id": "on-the-origin-of-cultural-biases-in-language-models-from-pre-training", "name": "On The Origin of Cultural Biases in Language Models: From Pre-training Data to Linguistic Phenomena", "type": "paper", "country": "INTL", "org": "tareknaous", "license": "unknown", "modality": "text", "tasks": [ "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2501.04662", "github": "https://github.com/tareknaous/camel2" }, "year": 2025, "venue": "arXiv 2025", "notes": "We introduce CAMeL-2, a parallel Arabic-English benchmark of 58,086 entities associated with Arab and Western cultures and 367 masked natural contexts.", "metrics": null }, { "id": "palm-a-culturally-inclusive-and-linguistically-diverse-dataset-for-ara", "name": "Palm: A Culturally Inclusive and Linguistically Diverse Dataset for Arabic LLMs", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "culture", "instructions", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2503.00151" }, "dialects": [ "msa", "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "PALM: community-built instruction dataset covering all 22 Arab countries in MSA and dialects to evaluate LLM cultural inclusivity.", "metrics": null }, { "id": "palmx-2025-the-first-shared-task-on-benchmarking-llms-on-arabic-and-is", "name": "PalmX 2025: The First Shared Task on Benchmarking LLMs on Arabic and Islamic Culture", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2509.02550" }, "dialects": [ "msa" ], "year": 2025, "venue": "arXiv 2025", "notes": "We introduce PalmX 2025, the first shared task designed to benchmark the cultural competence of LLMs in these specific domains.", "metrics": null }, { "id": "peach-a-sentence-aligned-parallel-english-arabic-corpus-for-healthcare", "name": "PEACH: A sentence-aligned Parallel English-Arabic Corpus for Healthcare", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2508.05722" }, "year": 2025, "venue": "arXiv 2025", "notes": "This paper introduces PEACH, a sentence-aligned parallel English-Arabic corpus of healthcare texts encompassing patient information leaflets.", "metrics": null }, { "id": "pearl-a-multimodal-culturally-aware-arabic-instruction-dataset", "name": "Pearl: A Multimodal Culturally-Aware Arabic Instruction Dataset", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "instruction-tuning", "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2505.21979" }, "year": 2025, "venue": "arXiv 2025", "notes": "To address this gap, we introduce PEARL, a large-scale Arabic multimodal dataset and benchmark explicitly designed for cultural understanding.", "metrics": null }, { "id": "phoneme-level-mispronunciation-detection-in-quranic-recitation-using-s", "name": "Phoneme-level mispronunciation detection in Quranic recitation using ShallowTransformer", "type": "paper", "country": "TN", "org": "University of Tunis", "license": "unknown", "modality": "text", "tasks": [ "quran", "pronunciation" ], "links": { "paper": "https://aclanthology.org/2025.arabicnlp-sharedtasks.63/" }, "year": 2025, "venue": "ArabicNLP 2025", "notes": "We present ShallowTransformer, a lightweight and computationally efficient transformer model leveraging Wav2vec2.0 features and trained with CTC loss.", "metrics": null }, { "id": "poem-meter-classification-of-recited-arabic-poetry-integrating-high-re", "name": "Poem Meter Classification of Recited Arabic Poetry: Integrating High-Resource Systems for a Low-Resource Task", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation", "poetry", "quran" ], "links": { "paper": "https://arxiv.org/abs/2504.12172" }, "year": 2025, "venue": "arXiv 2025", "notes": "We propose a state-of-the-art framework to identify the poem meters of recited Arabic poetry.", "metrics": null }, { "id": "proper-noun-diacritization-for-arabic-wikipedia-a-benchmark-dataset", "name": "Proper Noun Diacritization for Arabic Wikipedia: A Benchmark Dataset", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "diacritization", "ner", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2505.02656" }, "year": 2025, "venue": "arXiv 2025", "notes": "We introduce a new manually diacritized dataset of Arabic proper nouns of various origins with their English Wikipedia equivalent glosses.", "metrics": null }, { "id": "propxplain-can-llms-enable-explainable-propaganda-detection", "name": "PropXplain: Can LLMs Enable Explainable Propaganda Detection?", "type": "paper", "country": "INTL", "org": "firojalam", "license": "unknown", "modality": "text", "tasks": [ "nlp" ], "links": { "paper": "https://arxiv.org/abs/2502.16550", "github": "https://github.com/firojalam/PropXplain" }, "year": 2025, "venue": "arXiv 2025", "notes": "To address this issue, we propose a multilingual (i.e., Arabic and English) explanation-enhanced dataset, the first of its kind.", "metrics": null }, { "id": "qias-2025-overview-of-the-shared-task-on-islamic-inheritance-reasoning", "name": "QIAS 2025: Overview of the Shared Task on Islamic Inheritance Reasoning and Knowledge Assessment", "type": "paper", "country": "QA", "org": "Hamad Bin Khalifa University", "license": "unknown", "modality": "text", "tasks": [ "reasoning", "shared-task" ], "links": { "paper": "https://aclanthology.org/2025.arabicnlp-sharedtasks.117/" }, "year": 2025, "venue": "ArabicNLP 2025", "notes": "This paper provides a comprehensive overview of the QIAS 2025 shared task, organized as part of the ArabicNLP 2025 conference and co-located with EMNLP 2025.", "metrics": null }, { "id": "quranmorph-morphologically-annotated-quranic-corpus", "name": "QuranMorph: Morphologically Annotated Quranic Corpus", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "quran" ], "links": { "paper": "https://arxiv.org/abs/2506.18148", "website": "https://sina.birzeit.edu/quran" }, "dialects": [ "classical" ], "year": 2025, "venue": "arXiv 2025", "notes": "We present the QuranMorph corpus, a morphologically annotated corpus for the Quran (77,429 tokens).", "metrics": null }, { "id": "sage-spliced-audio-generated-data-for-enhancing-foundational-models-in", "name": "SAGE: Spliced-Audio Generated Data for Enhancing Foundational Models in Low-Resource Arabic-English Code-Switched Speech Recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2506.22143" }, "year": 2025, "venue": "arXiv 2025", "notes": "This paper investigates the performance of various speech SSL models on dialectal Arabic (DA) and Arabic-English code-switched (CS) speech.", "metrics": null }, { "id": "sard-a-large-scale-synthetic-arabic-ocr-dataset-for-book-style-text-re", "name": "SARD: A Large-Scale Synthetic Arabic OCR Dataset for Book-Style Text Recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "ocr", "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2505.24600" }, "year": 2025, "venue": "arXiv 2025", "notes": "To address this significant gap, we introduce SARD (Large-Scale Synthetic Arabic OCR Dataset).", "metrics": null }, { "id": "saudi-alignment-benchmark-assessing-llms-alignment-with-cultural-norms", "name": "Saudi-Alignment Benchmark: Assessing LLMs Alignment with Cultural Norms and Domain Knowledge in the Saudi Context", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://aclanthology.org/2025.arabicnlp-main.11/" }, "year": 2025, "venue": "ArabicNLP 2025", "notes": "To address this gap and the challenge LLMs face with non-Western cultural nuance, this study introduces the Saudi-Alignment Benchmark.", "metrics": null }, { "id": "senwave-a-fine-grained-multi-language-sentiment-analysis-dataset-sourc", "name": "SenWave: A Fine-Grained Multi-Language Sentiment Analysis Dataset Sourced from COVID-19 Tweets", "type": "paper", "country": "INTL", "org": "gitdevqiang", "license": "unknown", "modality": "text", "tasks": [ "chat", "translation", "sentiment", "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2510.08214", "github": "https://github.com/gitdevqiang/SenWave" }, "year": 2025, "venue": "arXiv 2025", "notes": "We introduce SenWave, a novel fine-grained multi-language sentiment analysis dataset specifically designed for analyzing COVID-19 tweets.", "metrics": null }, { "id": "shawarma-chats-a-benchmark-exact-dialogue-evaluation-platter-in-egypti", "name": "Shawarma Chats: A Benchmark Exact Dialogue & Evaluation Platter in Egyptian, Maghrebi & Modern Standard Arabic—A Triple-Dialect Feast for Hungry Language Models", "type": "paper", "country": "INTL", "org": "University of Siena", "license": "unknown", "modality": "text", "tasks": [ "dialogue", "evaluation" ], "links": { "paper": "https://aclanthology.org/2025.arabicnlp-main.39/" }, "dialects": [ "egy", "magh", "msa" ], "year": 2025, "venue": "ArabicNLP 2025", "notes": "Benchmark of 30,000 six-turn Wikipedia-grounded dialogues in Egyptian, Maghrebi and Modern Standard Arabic.", "metrics": null }, { "id": "tedxtn-a-three-way-speech-translation-corpus-for-code-switched-tunisia", "name": "TEDxTN: A Three-way Speech Translation Corpus for Code-Switched Tunisian Arabic - English", "type": "paper", "country": "TN", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "speech-translation", "asr" ], "links": { "paper": "https://arxiv.org/abs/2511.10780" }, "dialects": [ "magh" ], "year": 2025, "venue": "arXiv 2025", "notes": "TEDxTN: first public Tunisian Arabic to English speech translation corpus, 108 code-switched TEDx talks with about 25 hours of speech.", "metrics": null }, { "id": "tell-me-habibi-is-it-real-or-fake", "name": "Tell me Habibi, is it Real or Fake?", "type": "paper", "country": "INTL", "org": "kartik060702", "license": "unknown", "modality": "multimodal", "tasks": [ "tts", "evaluation", "speech", "vision" ], "links": { "paper": "https://arxiv.org/abs/2505.22581", "hf": "https://huggingface.co/datasets/kartik060702/ArEnAV-Full" }, "year": 2025, "venue": "arXiv 2025", "notes": "Deepfake generation methods are evolving fast, making fake media harder to detect and raising serious societal concerns.", "metrics": { "downloads": 32, "likes": 7, "lastModified": "2025-06-01" } }, { "id": "the-arageneval-shared-task-on-arabic-authorship-style-transfer-and-ai", "name": "The AraGenEval Shared Task on Arabic Authorship Style Transfer and AI Generated Text Detection", "type": "paper", "country": "SA", "org": "KFUPM", "license": "unknown", "modality": "text", "tasks": [ "shared-task" ], "links": { "paper": "https://aclanthology.org/2025.arabicnlp-sharedtasks.1/" }, "year": 2025, "venue": "ArabicNLP 2025", "notes": "We present an overview of the AraGenEval shared task, organized as part of the ArabicNLP 2025 conference.", "metrics": null }, { "id": "the-cross-lingual-cost-retrieval-biases-in-rag-over-arabic-english-cor", "name": "The Cross-Lingual Cost: Retrieval Biases in RAG over Arabic-English Corpora", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "translation", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2507.07543" }, "year": 2025, "venue": "arXiv 2025", "notes": "Finally, we propose two simple retrieval strategies that address this source of failure by enforcing equal retrieval from both languages.", "metrics": null }, { "id": "the-landscape-of-arabic-large-language-models", "name": "The Landscape of Arabic Large Language Models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "year": 2025, "venue": "arXiv 2025", "tasks": [ "survey", "llm" ], "links": { "paper": "https://arxiv.org/html/2506.01340v1" }, "notes": "Survey of Arabic LLMs: models, data, and evaluation to date.", "metrics": null }, { "id": "tool-calling-for-arabic-llms-data-strategies-and-instruction-tuning", "name": "Tool Calling for Arabic LLMs: Data Strategies and Instruction Tuning", "type": "paper", "country": "QA", "org": "QCRI", "license": "unknown", "modality": "text", "tasks": [ "instruction-tuning", "tool-calling" ], "links": { "paper": "https://aclanthology.org/2025.arabicnlp-main.28/" }, "year": 2025, "venue": "ArabicNLP 2025", "notes": "Tool calling is a critical capability that allows Large Language Models (LLMs) to interact with external systems, significantly expanding their utility.", "metrics": null }, { "id": "towards-a-unified-benchmark-for-arabic-pronunciation-assessment-qurani", "name": "Towards a Unified Benchmark for Arabic Pronunciation Assessment: Quranic Recitation as Case Study", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "evaluation", "quran" ], "links": { "paper": "https://arxiv.org/abs/2506.07722" }, "dialects": [ "classical", "msa" ], "year": 2025, "venue": "arXiv 2025", "notes": "We present a unified benchmark for mispronunciation detection in Modern Standard Arabic (MSA) using Qur'anic recitation as a case study.", "metrics": null }, { "id": "tunifra-a-tunisian-arabic-speech-corpus-with-orthographic-transcriptio", "name": "TuniFra: A Tunisian Arabic Speech Corpus with Orthographic Transcriptions and French Translations", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "translation" ], "links": { "paper": "https://aclanthology.org/2025.arabicnlp-main.5/" }, "dialects": [ "magh" ], "year": 2025, "venue": "ArabicNLP 2025", "notes": "We introduce TuniFra, a novel and comprehensive corpus developed to advance research in Automatic Speech Recognition (ASR) and Speech-to-Text Translation.", "metrics": null }, { "id": "voxlect-a-speech-foundation-model-benchmark-for-modeling-dialects-and", "name": "Voxlect: A Speech Foundation Model Benchmark for Modeling Dialects and Regional Languages Around the Globe", "type": "paper", "country": "INTL", "org": "tiantiaf0627", "license": "unknown", "modality": "speech", "tasks": [ "asr", "dialect-id", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2508.01691", "github": "https://github.com/tiantiaf0627/voxlect" }, "dialects": [ "mixed" ], "year": 2025, "venue": "arXiv 2025", "notes": "We present Voxlect, a novel benchmark for modeling dialects and regional languages worldwide using speech foundation models.", "metrics": null }, { "id": "wasm-a-pipeline-for-constructing-structured-arabic-interleaved-multimo", "name": "Wasm: A Pipeline for Constructing Structured Arabic Interleaved Multimodal Corpora", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "pretraining", "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2511.07080" }, "year": 2025, "venue": "arXiv 2025", "notes": "We present our pipeline Wasm for processing the Common Crawl dataset to create a new Arabic multimodal dataset that uniquely provides markdown output.", "metrics": null }, { "id": "zero-shot-and-fine-tuned-evaluation-of-generative-llms-for-arabic-word", "name": "Zero-Shot and Fine-Tuned Evaluation of Generative LLMs for Arabic Word Sense Disambiguation", "type": "paper", "country": "SD", "org": "University of Khartoum", "license": "unknown", "modality": "text", "tasks": [ "morphology" ], "links": { "paper": "https://aclanthology.org/2025.arabicnlp-main.24/" }, "year": 2025, "venue": "ArabicNLP 2025", "notes": "This paper benchmarks large generative language models (LLMs) for Arabic Word Sense Disambiguation (WSD) under both zero-shot and fine-tuning conditions.", "metrics": null }, { "id": "arabicmmlu-assessing-massive-multitask-language-understandin", "name": "ArabicMMLU: Assessing Massive Multitask Language Understanding in Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2402.12840" }, "year": 2024, "venue": "Annual Meeting of the Association for Computational Linguistics", "citations": 120, "notes": "The focus of language model evaluation has transitioned towards reasoning and knowledge-intensive tasks, driven by advancements in pretraining large models.", "metrics": null }, { "id": "real-time-arabic-sign-language-recognition-using-a-hybrid-de", "name": "Real-Time Arabic Sign Language Recognition Using a Hybrid Deep Learning Model", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "multimodal" ], "links": { "paper": "https://doi.org/10.3390/s24113683" }, "year": 2024, "venue": "Italian National Conference on Sensors", "citations": 67, "notes": "Sign language is an essential means of communication for individuals with hearing disabilities.", "metrics": null }, { "id": "darijabert-a-step-forward-in-nlp-for-the-written-moroccan-di", "name": "DarijaBERT: a step forward in NLP for the written Moroccan dialect", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://doi.org/10.1007/s41060-023-00498-2" }, "year": 2024, "venue": "International Journal of Data Science and Analysis", "citations": 58, "notes": "The established performance of existing transformer-based language models, delivering state-of-the-art results on numerous downstream tasks, is noteworthy.", "metrics": null }, { "id": "sentiment-analysis-of-arabic-social-media-texts-a-machine-le", "name": "Sentiment analysis of Arabic social media texts: A machine learning approach to deciphering customer perceptions", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1016/j.heliyon.2024.e27863" }, "year": 2024, "venue": "Heliyon", "citations": 53, "notes": "Sentiment analysis (SA) is a subfield of artificial intelligence that entails natural language processing.", "metrics": null }, { "id": "improving-neural-machine-translation-for-low-resource-langua", "name": "Improving neural machine translation for low resource languages through non-parallel corpora: a case study of Egyptian dialect to modern standard Arabic translation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "dataset" ], "links": { "paper": "https://doi.org/10.1038/s41598-023-51090-4" }, "year": 2024, "venue": "Scientific Reports", "citations": 46, "notes": "Machine translation for low-resource languages poses significant challenges, primarily due to the limited availability of data.", "metrics": null }, { "id": "sentiment-analysis-for-arabic-call-center-notes-using-machin", "name": "Sentiment analysis for Arabic call center notes using machine learning techniques", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.32629/jai.v7i3.940" }, "year": 2024, "venue": "Journal of Autonomous Intelligence", "citations": 44, "notes": "Call centers handle thousands of incoming calls daily, encompassing a diverse array of categories including product inquiries, complaints, and more.", "metrics": null }, { "id": "arabic-sarcasm-detection-an-enhanced-fine-tuned-language-mod", "name": "Arabic sarcasm detection: An enhanced fine-tuned language model approach", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "classification", "pretraining" ], "links": { "paper": "https://doi.org/10.1016/j.asej.2024.102736" }, "year": 2024, "venue": "Ain Shams Engineering Journal", "citations": 43, "metrics": null }, { "id": "deberta-bilstm-a-multi-label-classification-model-of-arabic", "name": "DeBERTa-BiLSTM: A multi-label classification model of Arabic medical questions using pre-trained models and deep learning", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa", "classification", "pretraining" ], "links": { "paper": "https://doi.org/10.1016/j.compbiomed.2024.107921" }, "year": 2024, "venue": "Comput. Biol. Medicine", "citations": 41, "notes": "It is wise to investigate past and present epidemics in the hopes of profiting from them and being better prepared for future ones.", "metrics": null }, { "id": "unveiling-the-new-frontier-chatgpt-3-powered-translation-for", "name": "Unveiling the New Frontier: ChatGPT-3 Powered Translation for Arabic-English Language Pairs", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "evaluation" ], "links": { "paper": "https://doi.org/10.17507/tpls.1402.05" }, "year": 2024, "venue": "Theory and Practice in Language Studies", "citations": 39, "notes": "This study evaluates the aptitude of ChatGPT for Arabic-English machine translation.", "metrics": null }, { "id": "empowering-communication-a-deep-learning-framework-for-arabi", "name": "Empowering Communication: A Deep Learning Framework for Arabic Sign Language Recognition with an Attention Mechanism", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "multimodal" ], "links": { "paper": "https://doi.org/10.3390/computers13060153" }, "year": 2024, "venue": "De Computis", "citations": 38, "notes": "This article emphasises the urgent need for appropriate communication tools for communities of people who are deaf or hard-of-hearing, with a specific emphasis", "metrics": null }, { "id": "from-text-to-insight-an-integrated-cnn-bilstm-gru-model-for", "name": "From Text to Insight: An Integrated CNN-BiLSTM-GRU Model for Arabic Cyberbullying Detection", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2024.3431939" }, "year": 2024, "venue": "IEEE Access", "citations": 38, "notes": "Several research on cyberbullying detection have employed different deep learning and machine learning methodologies to achieve promising outcomes.", "metrics": null }, { "id": "adocrnet-a-deep-learning-ocr-for-arabic-documents-recognitio", "name": "ADOCRNet: A Deep Learning OCR for Arabic Documents Recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "ocr" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2024.3379530" }, "year": 2024, "venue": "IEEE Access", "citations": 37, "notes": "In recent years, Optical character recognition (OCR) has experienced a resurgence of interest especially for contemporary Arabic data.", "metrics": null }, { "id": "arabic-sentiment-analysis-of-monkeypox-using-deep-neural-net", "name": "Arabic sentiment analysis of Monkeypox using deep neural network and optimized hyperparameters of machine learning algorithms", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1007/s13278-023-01188-4" }, "year": 2024, "venue": "Social Network Analysis and Mining", "citations": 37, "metrics": null }, { "id": "advancements-and-challenges-in-arabic-sentiment-analysis-a-d", "name": "Advancements and challenges in Arabic sentiment analysis: A decade of methodologies, applications, and resource development", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1016/j.heliyon.2024.e39786" }, "year": 2024, "venue": "Heliyon", "citations": 36, "notes": "The exponential growth of digital information, particularly user-generated content on social media and blogging platforms, has underscored the importance of", "metrics": null }, { "id": "language-discrepancies-in-the-performance-of-generative-arti", "name": "Language discrepancies in the performance of generative artificial intelligence models: an examination of infectious disease queries in English and Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://doi.org/10.1186/s12879-024-09725-y" }, "year": 2024, "venue": "BMC Infectious Diseases", "citations": 35, "notes": "This study aimed to compare AI model efficiency in English and Arabic for infectious disease queries.", "metrics": null }, { "id": "translating-classical-arabic-verse-human-translation-vs-ai-l", "name": "Translating classical Arabic verse: human translation vs. AI large language models (Gemini and ChatGPT)", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "pretraining", "evaluation" ], "links": { "paper": "https://doi.org/10.1080/23311886.2024.2410998" }, "year": 2024, "venue": "Cogent Social Sciences", "citations": 35, "notes": "This paper examines the translation of 15 individual Classical Arabic verses by comparing the English renditions provided by a human translator and two", "metrics": null }, { "id": "sada-saudi-audio-dataset-for-arabic", "name": "SADA: Saudi Audio Dataset for Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "dataset" ], "links": { "paper": "https://doi.org/10.1109/ICASSP48485.2024.10446243" }, "year": 2024, "venue": "IEEE International Conference on Acoustics, Speech, and Signal Processing", "citations": 34, "notes": "This paper introduces SADA, the Saudi Audio Dataset for Arabic, with 668 hours of high-quality audio suitable for supervised training.", "metrics": null }, { "id": "ai-generated-text-detector-for-arabic-language-using-encoder", "name": "AI-Generated Text Detector for Arabic Language Using Encoder-Based Transformer Architecture", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification" ], "links": { "paper": "https://doi.org/10.3390/bdcc8030032" }, "year": 2024, "venue": "Big Data and Cognitive Computing", "citations": 32, "notes": "This study introduces a novel AI text classifier designed specifically for Arabic, tackling the distinct challenges inherent in processing this language.", "metrics": null }, { "id": "user-satisfaction-with-arabic-covid-19-apps-sentiment-analys", "name": "User satisfaction with Arabic COVID-19 apps: Sentiment analysis of users' reviews using machine learning techniques", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1016/j.ipm.2024.103644" }, "year": 2024, "venue": "Information Processing & Management", "citations": 32, "metrics": null }, { "id": "arabic-fake-news-detection-in-social-media-context-using-wor", "name": "Arabic Fake News Detection in Social Media Context Using Word Embeddings and Pre-trained Transformers", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "classification", "pretraining" ], "links": { "paper": "https://doi.org/10.1007/s13369-024-08959-x" }, "year": 2024, "venue": "The Arabian journal for science and engineering", "citations": 30, "metrics": null }, { "id": "hate-speech-detection-with-adhar-a-multi-dialectal-hate-spee", "name": "Hate speech detection with ADHAR: a multi-dialectal hate speech corpus in Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification", "dataset" ], "links": { "paper": "https://doi.org/10.3389/frai.2024.1391472" }, "year": 2024, "venue": "Frontiers Artif. Intell.", "citations": 30, "notes": "Hate speech detection in Arabic poses a complex challenge due to the dialectal diversity across the Arab world.", "metrics": null }, { "id": "a-tinydl-model-for-gesture-based-air-handwriting-arabic-numb", "name": "A TinyDL Model for Gesture-Based Air Handwriting Arabic Numbers and Simple Arabic Letters Recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "ocr" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2024.3406631" }, "year": 2024, "venue": "IEEE Access", "citations": 29, "notes": "The application of tiny machine learning (TinyML) in human-computer interaction is revolutionizing gesture recognition technologies.", "metrics": null }, { "id": "arabic-sentiment-analysis-for-chatgpt-using-machine-learning", "name": "Arabic Sentiment Analysis for ChatGPT Using Machine Learning Classification Algorithms: A Hyperparameter Optimization Technique", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "classification", "evaluation" ], "links": { "paper": "https://doi.org/10.1145/3638285" }, "year": 2024, "venue": "ACM Trans. Asian Low Resour. Lang. Inf. Process.", "citations": 29, "notes": "This study centers on ChatGPT, a popular machine learning model engaging in dialogues with users, garnering attention for its exceptional performance and", "metrics": null }, { "id": "systematic-review-of-english-arabic-machine-translation-post", "name": "Systematic Review of English/Arabic Machine Translation Postediting: Implications for AI Application in Translation Research and Pedagogy", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "translation" ], "links": { "paper": "https://doi.org/10.3390/informatics11020023" }, "year": 2024, "venue": "Informatics", "citations": 29, "notes": "The twenty-first century has witnessed an extensive evolution in translation practice thanks to the accelerated progress in machine translation tools and", "metrics": null }, { "id": "the-performance-of-openai-chatgpt-4-and-google-gemini-in-vir", "name": "The performance of OpenAI ChatGPT-4 and Google Gemini in virology multiple-choice questions: a comparative analysis of English and Arabic responses", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa", "evaluation" ], "links": { "paper": "https://doi.org/10.1186/s13104-024-06920-7" }, "year": 2024, "venue": "BMC Research Notes", "citations": 29, "notes": "The integration of artificial intelligence (AI) in healthcare education is inevitable.", "metrics": null }, { "id": "bert-based-model-for-aspect-based-sentiment-analysis-for-ana", "name": "BERT-Based Model for Aspect-Based Sentiment Analysis for Analyzing Arabic Open-Ended Survey Responses: A Case Study", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "sentiment", "pretraining" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2023.3348342" }, "year": 2024, "venue": "IEEE Access", "citations": 28, "notes": "Educational institutions typically gather feedback from beneficiaries through formal surveys.", "metrics": null }, { "id": "dallah-a-dialect-aware-multimodal-large-language-model-for-a", "name": "Dallah: A Dialect-Aware Multimodal Large Language Model for Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "multimodal", "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2407.18129" }, "year": 2024, "venue": "ARABICNLP", "citations": 28, "notes": "Recent advancements have significantly enhanced the capabilities of Multimodal Large Language Models (MLLMs) in generating and understanding image-to-text", "metrics": null }, { "id": "detection-of-arabic-offensive-language-in-social-media-using", "name": "Detection of Arabic offensive language in social media using machine learning models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification" ], "links": { "paper": "https://doi.org/10.1016/j.iswa.2024.200376" }, "year": 2024, "venue": "Intelligent Systems with Applications", "citations": 28, "metrics": null }, { "id": "arabic-sign-language-letters-recognition-using-vision-transf", "name": "Arabic sign language letters recognition using Vision Transformer", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "multimodal" ], "links": { "paper": "https://doi.org/10.1007/s11042-024-18681-3" }, "year": 2024, "venue": "Multimedia tools and applications", "citations": 27, "metrics": null }, { "id": "evaluating-chatgpt-performance-in-arabic-dialects-a-comparat", "name": "Evaluating ChatGPT performance in Arabic dialects: A comparative study showing defects in responding to Jordanian and Tunisian general health prompts", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://doi.org/10.58496/mjaih/2024/001" }, "year": 2024, "venue": "Mesopotamian Journal of Artificial Intelligence in Healthcare", "citations": 27, "notes": "Background: The role of artificial intelligence (AI) is increasingly recognized to enhance digital health literacy.", "metrics": null }, { "id": "linguistic-feature-fusion-for-arabic-fake-news-detection-and", "name": "Linguistic feature fusion for Arabic fake news detection and named entity recognition using reinforcement learning and swarm optimization", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "ner", "classification" ], "links": { "paper": "https://doi.org/10.1016/j.neucom.2024.128078" }, "year": 2024, "venue": "Neurocomputing", "citations": 26, "metrics": null }, { "id": "a-tinyml-model-for-gesture-based-air-handwriting-arabic-numb", "name": "A TinyML Model for Gesture-Based Air Handwriting Arabic Numbers Recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "ocr" ], "links": { "paper": "https://doi.org/10.1016/j.procs.2024.05.070" }, "year": 2024, "venue": "Procedia Computer Science", "citations": 25, "metrics": null }, { "id": "enhancing-arabic-dialect-detection-on-social-media-a-hybrid", "name": "Enhancing Arabic Dialect Detection on Social Media: A Hybrid Model with an Attention Mechanism", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id", "classification" ], "links": { "paper": "https://doi.org/10.3390/info15060316" }, "year": 2024, "venue": "Inf.", "citations": 25, "notes": "Recently, the widespread use of social media and easy access to the Internet have brought about a significant transformation in the type of textual data", "metrics": null }, { "id": "hatformer-historic-handwritten-arabic-text-recognition-with", "name": "HATFormer: Historic Handwritten Arabic Text Recognition with Transformers", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "ocr" ], "links": { "paper": "https://arxiv.org/abs/2410.02179" }, "year": 2024, "venue": "arXiv.org", "citations": 25, "notes": "Arabic handwritten text recognition (HTR) is challenging, especially for historical texts, due to diverse writing styles and the intrinsic features of Arabic", "metrics": null }, { "id": "a-comprehensive-analysis-of-various-tokenizers-for-arabic-la", "name": "A Comprehensive Analysis of Various Tokenizers for Arabic Large Language Models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "preprocessing", "pretraining" ], "links": { "paper": "https://doi.org/10.3390/app14135696" }, "year": 2024, "venue": "Applied Sciences", "citations": 24, "notes": "Pretrained language models have achieved great success in various natural language understanding (NLU) tasks due to their capacity to capture deep", "metrics": null }, { "id": "arabbert-lstm-improving-arabic-sentiment-analysis-based-on-t", "name": "ArabBert-LSTM: improving Arabic sentiment analysis based on transformer model and Long Short-Term Memory", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.3389/frai.2024.1408845" }, "year": 2024, "venue": "Frontiers Artif. Intell.", "citations": 24, "notes": "Sentiment analysis also referred to as opinion mining, plays a significant role in automating the identification of negative, positive, or neutral sentiments", "metrics": null }, { "id": "efhamni-a-deep-learning-based-saudi-sign-language-recognitio", "name": "Efhamni: A Deep Learning-Based Saudi Sign Language Recognition Application", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "multimodal" ], "links": { "paper": "https://doi.org/10.3390/s24103112" }, "year": 2024, "venue": "Italian National Conference on Sensors", "citations": 24, "notes": "Deaf and hard-of-hearing people mainly communicate using sign language, which is a set of signs made using hand gestures combined with facial expressions to", "metrics": null }, { "id": "a-machine-learning-approach-to-cyberbullying-detection-in-ar", "name": "A Machine Learning Approach to Cyberbullying Detection in Arabic Tweets", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification" ], "links": { "paper": "https://doi.org/10.32604/cmc.2024.048003" }, "year": 2024, "venue": "Comput. Mater. Continua", "citations": 23, "notes": "With the rapid growth of internet usage, a new situation has been created that enables practicing bullying.", "metrics": null }, { "id": "error-analysis-of-pretrained-language-models-plms-in-english", "name": "Error Analysis of Pretrained Language Models (PLMs) in English-to-Arabic Machine Translation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "pretraining" ], "links": { "paper": "https://doi.org/10.1007/s44230-024-00061-7" }, "year": 2024, "venue": "Human-Centric Intelligent Systems", "citations": 23, "notes": "Advances in neural machine translation utilizing pretrained language models (PLMs) have shown promise in improving the translation quality between diverse", "metrics": null }, { "id": "multiscale-cascaded-domain-based-approach-for-arabic-fake-re", "name": "Multiscale cascaded domain-based approach for Arabic fake reviews detection in e-commerce platforms", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification" ], "links": { "paper": "https://doi.org/10.1016/j.jksuci.2024.101926" }, "year": 2024, "venue": "Journal of King Saud University: Computer and Information Sciences", "citations": 23, "notes": "This study addresses this gap by introducing a full-gold standard dataset, the", "metrics": null }, { "id": "toward-robust-arabic-ai-generated-text-detection-tackling-di", "name": "Toward Robust Arabic AI-Generated Text Detection: Tackling Diacritics Challenges", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "diacritization", "classification" ], "links": { "paper": "https://doi.org/10.3390/info15070419" }, "year": 2024, "venue": "Inf.", "citations": 23, "notes": "This study introduces robust Arabic text detection models using Transformer-based", "metrics": null }, { "id": "aratrust-an-evaluation-of-trustworthiness-for-llms-in-arabic", "name": "AraTrust: An Evaluation of Trustworthiness for LLMs in Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2403.09017" }, "year": 2024, "venue": "International Conference on Computational Linguistics", "citations": 22, "notes": "The swift progress and widespread acceptance of artificial intelligence (AI) systems highlight a pressing requirement to comprehend both the capabilities and", "metrics": null }, { "id": "enhancedbert-a-feature-rich-ensemble-model-for-arabic-word-s", "name": "EnhancedBERT: A feature-rich ensemble model for Arabic word sense disambiguation with statistical analysis and optimized data collection", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "morphology" ], "links": { "paper": "https://doi.org/10.1016/j.jksuci.2023.101911" }, "year": 2024, "venue": "Journal of King Saud University: Computer and Information Sciences", "citations": 22, "notes": "Accurate assignment of meaning to a word based on its context, known as Word Sense Disambiguation (WSD), remains challenging across languages.", "metrics": null }, { "id": "evaluating-arabic-emotion-recognition-task-using-chatgpt-mod", "name": "Evaluating Arabic Emotion Recognition Task Using ChatGPT Models: A Comparative Analysis between Emotional Stimuli Prompt, Fine-Tuning, and In-Context Learning", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "evaluation" ], "links": { "paper": "https://doi.org/10.3390/jtaer19020058" }, "year": 2024, "venue": "Journal of Theoretical and Applied Electronic Commerce Research", "citations": 22, "notes": "Textual emotion recognition (TER) has significant commercial potential since it can be used as an excellent tool to monitor a brand/business reputation", "metrics": null }, { "id": "arabic-speech-recognition-advancement-and-challenges", "name": "Arabic Speech Recognition: Advancement and Challenges", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2024.3376237" }, "year": 2024, "venue": "IEEE Access", "citations": 21, "notes": "Speech recognition is a captivating process that revolutionizes human-computer interactions, allowing us to interact and control machines through spoken", "metrics": null }, { "id": "cnn-based-methods-for-offline-arabic-handwriting-recognition", "name": "CNN-based Methods for Offline Arabic Handwriting Recognition: A Review", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "survey", "ocr" ], "links": { "paper": "https://doi.org/10.1007/s11063-024-11544-w" }, "year": 2024, "venue": "Neural Processing Letters", "citations": 21, "notes": "Arabic Handwriting Recognition (AHR) is a complex task involving the transformation of handwritten Arabic text from image format into machine-readable data", "metrics": null }, { "id": "tayseer-a-novel-ai-powered-arabic-chatbot-framework-for-tech", "name": "Tayseer: A Novel AI-Powered Arabic Chatbot Framework for Technical and Vocational Student Helpdesk Services and Enhancing Student Interactions", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "benchmark", "dataset" ], "links": { "paper": "https://doi.org/10.3390/app14062547" }, "year": 2024, "venue": "Applied Sciences", "citations": 21, "notes": "The rise of conversational agents (CAs) like chatbots in education has increased the demand for advisory services.", "metrics": null }, { "id": "a-brief-review-on-preprocessing-text-in-arabic-language-data", "name": "A Brief Review on Preprocessing Text in Arabic Language Dataset: Techniques and Challenges", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "preprocessing", "dataset" ], "links": { "paper": "https://doi.org/10.58496/bjai/2024/007" }, "year": 2024, "venue": "Babylonian Journal of Artificial Intelligence", "citations": 20, "notes": "The preprocessing of Arabic text still presents unique challenges due to the language's rich morphology, complex grammar, and", "metrics": null }, { "id": "breaking-language-barriers-with-chatgpt-enhancing-low-resour", "name": "Breaking language barriers with ChatGPT: enhancing low-resource machine translation between algerian arabic and MSA", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "evaluation" ], "links": { "paper": "https://doi.org/10.1007/s41870-024-01926-7" }, "year": 2024, "venue": "International journal of information technology", "citations": 20, "metrics": null }, { "id": "embedding-search-for-quranic-texts-based-on-large-language-m", "name": "Embedding search for quranic texts based on large language models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "quran", "embedding", "pretraining" ], "links": { "paper": "https://doi.org/10.34028/21/2/7" }, "year": 2024, "venue": "˜The œinternational Arab journal of information technology", "citations": 20, "notes": "This paper is introduced in order to explore the use of large language models for semantic search of Quranic texts.", "metrics": null }, { "id": "neural-multi-task-learning-for-end-to-end-arabic-aspect-base", "name": "Neural multi-task learning for end-to-end Arabic aspect-based sentiment analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1016/j.csl.2024.101683" }, "year": 2024, "venue": "Computer Speech and Language", "citations": 20, "metrics": null }, { "id": "word-embedding-as-a-semantic-feature-extraction-technique-in", "name": "Word embedding as a semantic feature extraction technique in arabic natural language processing: an overview", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "embedding" ], "links": { "paper": "https://doi.org/10.34028/21/2/13" }, "year": 2024, "venue": "˜The œinternational Arab journal of information technology", "citations": 20, "notes": "Feature extraction has transformed the field of Natural Language Processing (NLP) by providing an effective way to represent linguistic features.", "metrics": null }, { "id": "a-comprehensive-review-on-arabic-offensive-language-and-hate", "name": "A comprehensive review on Arabic offensive language and hate speech detection on social media: methods, challenges and solutions", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "hate-speech", "classification" ], "links": { "paper": "https://doi.org/10.1007/s13278-024-01258-1" }, "year": 2024, "venue": "Social Network Analysis and Mining", "citations": 19, "metrics": null }, { "id": "alignment-at-pre-training-towards-native-alignment-for-arabi", "name": "Alignment at Pre-training! Towards Native Alignment for Arabic LLMs", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2412.03253" }, "year": 2024, "venue": "Neural Information Processing Systems", "citations": 19, "notes": "Traditional approaches focus on aligning models during the instruction tuning or reinforcement learning stages, referred to in this paper as `post alignment'.", "metrics": null }, { "id": "aracovtexfinder-leveraging-the-transformer-based-language-mo", "name": "AraCovTexFinder: Leveraging the transformer-based language model for Arabic COVID-19 text identification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://doi.org/10.1016/j.engappai.2024.107987" }, "year": 2024, "venue": "Engineering applications of artificial intelligence", "citations": 19, "metrics": null }, { "id": "code-mixing-unveiled-enhancing-the-hate-speech-detection-in", "name": "Code-mixing unveiled: Enhancing the hate speech detection in Arabic dialect tweets using machine learning models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification" ], "links": { "paper": "https://doi.org/10.1371/journal.pone.0305657" }, "year": 2024, "venue": "PLoS ONE", "citations": 19, "notes": "Technological developments over the past few decades have changed the way people communicate, with platforms like social media and blogs becoming vital channels", "metrics": null }, { "id": "exploring-retrieval-augmented-generation-in-arabic", "name": "Exploring Retrieval Augmented Generation in Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining", "benchmark", "dataset" ], "links": { "paper": "https://arxiv.org/abs/2408.07425" }, "year": 2024, "venue": "International Conference on Arabic Computational Linguistics", "citations": 19, "notes": "Recently, Retrieval Augmented Generation (RAG) has emerged as a powerful technique in natural language processing, combining the strengths of retrieval-based", "metrics": null }, { "id": "low-resource-arabic-dialects-transformer-neural-machine-tran", "name": "Low Resource Arabic Dialects Transformer Neural Machine Translation Improvement through Incremental Transfer of Shared Linguistic Features", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation" ], "links": { "paper": "https://doi.org/10.1007/s13369-023-08543-9" }, "year": 2024, "venue": "The Arabian journal for science and engineering", "citations": 19, "metrics": null }, { "id": "natural-language-processing-for-arabic-sentiment-analysis-a", "name": "Natural Language Processing for Arabic Sentiment Analysis: A Systematic Literature Review", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "sentiment" ], "links": { "paper": "https://doi.org/10.1109/TBDATA.2024.3366083" }, "year": 2024, "venue": "IEEE Transactions on Big Data", "citations": 19, "notes": "Sentiment analysis involves using computational methods to identify and classify opinions expressed in text, with the goal of determining whether the writer's", "metrics": null }, { "id": "novel-approach-for-arabic-fake-news-classification-using-emb", "name": "Novel approach for Arabic fake news classification using embedding from large language features with CNN-LSTM ensemble model and explainable AI", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "classification", "pretraining" ], "links": { "paper": "https://doi.org/10.1038/s41598-024-82111-5" }, "year": 2024, "venue": "Scientific Reports", "citations": 19, "notes": "This study addresses this critical issue by advancing fake news detection in Arabic and overcoming limitations in existing approaches.", "metrics": null }, { "id": "optimizing-large-language-models-for-arabic-healthcare-commu", "name": "Optimizing Large Language Models for Arabic Healthcare Communication: A Focus on Patient-Centered NLP Applications", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://doi.org/10.3390/bdcc8110157" }, "year": 2024, "venue": "Big Data and Cognitive Computing", "citations": 19, "notes": "Recent studies have highlighted the growing integration of Natural Language Processing (NLP) techniques and Large Language Models (LLMs) in healthcare.", "metrics": null }, { "id": "al-qasida-analyzing-llm-quality-and-accuracy-systematically", "name": "AL-QASIDA: Analyzing LLM Quality and Accuracy Systematically in Dialectal Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2412.04193" }, "year": 2024, "venue": "arXiv.org", "citations": 18, "notes": "Dialectal Arabic (DA) varieties are under-served by language technologies, particularly large language models (LLMs).", "metrics": null }, { "id": "an-arabic-visual-speech-recognition-framework-with-cnn-and-v", "name": "An arabic visual speech recognition framework with CNN and vision transformers for lipreading", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "asr", "multimodal" ], "links": { "paper": "https://doi.org/10.1007/s11042-024-18237-5" }, "year": 2024, "venue": "Multimedia tools and applications", "citations": 18, "metrics": null }, { "id": "arabic-automatic-speech-recognition-challenges-and-progress", "name": "Arabic Automatic Speech Recognition: Challenges and Progress", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr" ], "links": { "paper": "https://doi.org/10.1016/j.specom.2024.103110" }, "year": 2024, "venue": "Speech Communication", "citations": 18, "metrics": null }, { "id": "arabic-synonym-bert-based-adversarial-examples-for-text-clas", "name": "Arabic Synonym BERT-based Adversarial Examples for Text Classification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification", "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2402.03477" }, "year": 2024, "venue": "Conference of the European Chapter of the Association for Computational Linguistics", "citations": 18, "notes": "Text classification systems have been proven vulnerable to adversarial text examples, modified versions of the original text examples that are often unnoticed", "metrics": null }, { "id": "deep-learning-ensemble-and-supervised-machine-learning-for-a", "name": "Deep Learning, Ensemble and Supervised Machine Learning for Arabic Speech Emotion Recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.48084/etasr.7134" }, "year": 2024, "venue": "Engineering, Technology & Applied Science Research", "citations": 18, "notes": "Today, automatic emotion recognition in speech is one of the most important areas of research in signal processing.", "metrics": null }, { "id": "exploring-challenges-in-audiovisual-translation-a-comparativ", "name": "Exploring challenges in audiovisual translation: A comparative analysis of human- and AI-generated Arabic subtitles in Birdman", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "translation", "evaluation" ], "links": { "paper": "https://doi.org/10.1371/journal.pone.0311020" }, "year": 2024, "venue": "PLoS ONE", "citations": 18, "notes": "Movies often use allusions to add depth, create connections, and enrich the storytelling.", "metrics": null }, { "id": "optimised-cnn-architectures-for-handwritten-arabic-character", "name": "Optimised CNN Architectures for Handwritten Arabic Character Recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "ocr" ], "links": { "paper": "https://doi.org/10.32604/cmc.2024.052016" }, "year": 2024, "venue": "Computers, Materials & Continua", "citations": 18, "notes": "Handwritten character recognition is considered challenging compared with machine-printed characters due to the different human writing styles.", "metrics": null }, { "id": "stanceeval-2024-the-first-arabic-stance-detection-shared-tas", "name": "StanceEval 2024: The First Arabic Stance Detection Shared Task", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "classification", "evaluation" ], "links": { "paper": "https://aclanthology.org/2024.arabicnlp-1.88/" }, "year": 2024, "venue": "ARABICNLP", "citations": 18, "notes": "Recently, there has been a growing interest in analyzing user-generated text to understand opinions expressed on social media.", "metrics": null }, { "id": "a-bidirectional-arabic-sign-language-framework-using-deep-le", "name": "A Bidirectional Arabic Sign Language Framework Using Deep Learning and Fuzzy Matching Score", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "multimodal" ], "links": { "paper": "https://doi.org/10.3390/math12081155" }, "year": 2024, "venue": "Mathematics", "citations": 17, "notes": "Sign language is widely used to facilitate the communication process between deaf people and their surrounding environment.", "metrics": null }, { "id": "a-systematic-literature-review-of-hate-speech-identification", "name": "A systematic literature review of hate speech identification on Arabic Twitter data: research challenges and future directions", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "hate-speech" ], "links": { "paper": "https://doi.org/10.7717/peerj-cs.1966" }, "year": 2024, "venue": "PeerJ Computer Science", "citations": 17, "notes": "The automatic speech identification in Arabic tweets has generated substantial attention among academics in the fields of text mining and natural language", "metrics": null }, { "id": "arabic-dialect-identification-in-social-media-a-hybrid-model", "name": "Arabic dialect identification in social media: A hybrid model with transformer models and BiLSTM", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id" ], "links": { "paper": "https://doi.org/10.1016/j.heliyon.2024.e36280" }, "year": 2024, "venue": "Heliyon", "citations": 17, "notes": "Therefore, this study aims to use transformers to address the issue of ADI on social media.", "metrics": null }, { "id": "bert-based-arabic-diacritization-a-state-of-the-art-approach", "name": "BERT-Based Arabic Diacritization: A state-of-the-art approach for improving text accuracy and pronunciation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "diacritization", "pretraining" ], "links": { "paper": "https://doi.org/10.1016/j.eswa.2024.123416" }, "year": 2024, "venue": "Expert systems with applications", "citations": 17, "metrics": null }, { "id": "artificial-intelligence-generated-arabic-subtitles-insights", "name": "Artificial intelligence-generated Arabic subtitles: insights from Veed.io’s automatic speech recognition system of Jordanian Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "translation" ], "links": { "paper": "https://doi.org/10.1590/1983-3652.2024.46952" }, "year": 2024, "venue": "Texto Livre", "citations": 16, "notes": "This paper examines the errors that the automatic speech recognition (ASR) system of Veed.io produces when transcribing utterances spoken in Jordanian Arabic", "metrics": null }, { "id": "enhancing-arabic-sentiment-analysis-of-consumer-reviews-mach", "name": "Enhancing Arabic Sentiment Analysis of Consumer Reviews: Machine Learning and Deep Learning Methods Based on NLP", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.3390/a17110495" }, "year": 2024, "venue": "Algorithms", "citations": 16, "notes": "Sentiment analysis utilizes Natural Language Processing (NLP) techniques to extract opinions from text, which is critical for businesses looking to refine", "metrics": null }, { "id": "integrated-multi-layer-perceptron-neural-network-and-novel-f", "name": "Integrated multi-layer perceptron neural network and novel feature extraction for handwritten Arabic recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "ocr" ], "links": { "paper": "https://doi.org/10.5267/j.ijdns.2024.3.015" }, "year": 2024, "venue": "International Journal of Data and Network Science", "citations": 16, "notes": "Arabic handwritten script recognition presents an energetic area of study.", "metrics": null }, { "id": "pre-trained-language-model-ensemble-for-arabic-fake-news-det", "name": "Pre-Trained Language Model Ensemble for Arabic Fake News Detection", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification", "pretraining" ], "links": { "paper": "https://doi.org/10.3390/math12182941" }, "year": 2024, "venue": "Mathematics", "citations": 16, "notes": "Fake news detection (FND) remains a challenge due to its vast and varied sources, especially on social media platforms.", "metrics": null }, { "id": "a-hybrid-combination-of-cnn-attention-with-optimized-random", "name": "A hybrid combination of CNN Attention with optimized random forest with grey wolf optimizer to discriminate between Arabic hateful, abusive tweets", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech" ], "links": { "paper": "https://doi.org/10.1016/j.jksuci.2024.101961" }, "year": 2024, "venue": "Journal of King Saud University: Computer and Information Sciences", "citations": 15, "notes": "Arabic hateful speech recognition has long been a major area of focus in Natural Language Processing (NLP) research.", "metrics": null }, { "id": "a-robust-classification-approach-to-enhance-clinic-identific", "name": "A robust classification approach to enhance clinic identification from Arabic health text", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification" ], "links": { "paper": "https://doi.org/10.1007/s00521-024-09453-z" }, "year": 2024, "venue": "Neural computing & applications (Print)", "citations": 15, "metrics": null }, { "id": "an-analysis-of-customer-perception-using-lexicon-based-senti", "name": "An analysis of customer perception using lexicon-based sentiment analysis of Arabic Texts framework", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "dataset" ], "links": { "paper": "https://doi.org/10.1016/j.heliyon.2024.e30320" }, "year": 2024, "venue": "Heliyon", "citations": 15, "notes": "Sentiment Analysis (SA) employing Natural Language Processing (NLP) is pivotal in determining the positivity and negativity of customer feedback.", "metrics": null }, { "id": "enhancing-arabic-cyberbullying-detection-with-end-to-end-tra", "name": "Enhancing Arabic Cyberbullying Detection with End-to-End Transformer Model", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification" ], "links": { "paper": "https://doi.org/10.32604/cmes.2024.052291" }, "year": 2024, "venue": "Computer Modeling in Engineering & Sciences", "citations": 15, "notes": "To tackle this challenge, our study introduces a new approach employing Bidirectional Encoder Representations from the", "metrics": null }, { "id": "enhancing-communication-deep-learning-for-arabic-sign-langua", "name": "Enhancing communication: Deep learning for Arabic sign language translation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "translation", "multimodal" ], "links": { "paper": "https://doi.org/10.1515/eng-2024-0025" }, "year": 2024, "venue": "Open Engineering", "citations": 15, "notes": "This study explores the field of sign language recognition through machine learning, focusing on the development and comparative evaluation of various", "metrics": null }, { "id": "enhancing-semantic-similarity-understanding-in-arabic-nlp-wi", "name": "Enhancing Semantic Similarity Understanding in Arabic NLP with Nested Embedding Learning", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding" ], "links": { "paper": "https://arxiv.org/abs/2407.21139" }, "year": 2024, "venue": "arXiv.org", "citations": 15, "notes": "This work presents a novel framework for training Arabic nested embedding models through Matryoshka Embedding Learning, leveraging multilingual", "metrics": null }, { "id": "machine-learning-approach-for-arabic-handwritten-recognition", "name": "Machine Learning Approach for Arabic Handwritten Recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "ocr" ], "links": { "paper": "https://doi.org/10.3390/app14199020" }, "year": 2024, "venue": "Applied Sciences", "citations": 15, "notes": "Text recognition is an important area of the pattern recognition field.", "metrics": null }, { "id": "mtl-arabert-an-enhanced-multi-task-learning-model-for-arabic", "name": "MTL-AraBERT: An Enhanced Multi-Task Learning Model for Arabic Aspect-Based Sentiment Analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.3390/computers13040098" }, "year": 2024, "venue": "De Computis", "citations": 15, "notes": "Aspect-based sentiment analysis (ABSA) is a fine-grained type of sentiment analysis; it works on an aspect level.", "metrics": null }, { "id": "sentiment-analysis-of-imbalanced-arabic-data-using-sampling", "name": "Sentiment analysis of imbalanced Arabic data using sampling techniques and classification algorithms", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "classification" ], "links": { "paper": "https://doi.org/10.11591/eei.v13i1.5886" }, "year": 2024, "venue": "Bulletin of Electrical Engineering and Informatics", "citations": 15, "notes": "Sentiment analysis is a popular natural language processing task that recognizes the opinions or feelings of a piece of text.", "metrics": null }, { "id": "transformer-models-in-education-summarizing-science-textbook", "name": "Transformer Models in Education: Summarizing Science Textbooks with AraBART, MT5, AraT5, and mBART", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "summarization" ], "links": { "paper": "https://arxiv.org/abs/2406.07692" }, "year": 2024, "venue": "arXiv.org", "citations": 15, "notes": "Given this challenge, we have developed an advanced text summarization system targeting Arabic textbooks.", "metrics": null }, { "id": "101-billion-arabic-words-dataset", "name": "101 Billion Arabic Words Dataset", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2405.01590" }, "year": 2024, "venue": "arXiv 2024", "notes": "In recent years, Large Language Models have revolutionized the field of natural language processing.", "metrics": null }, { "id": "a-context-contrastive-inference-approach-to-partial-diacritization", "name": "A Context-Contrastive Inference Approach To Partial Diacritization", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "diacritization" ], "links": { "paper": "https://aclanthology.org/2024.arabicnlp-1.8/" }, "year": 2024, "venue": "ArabicNLP 2024", "notes": "In this light, we introduce Context-Contrastive Partial Diacritization (‘CCPD‘)—a novel approach to ‘PD‘ which integrates seamlessly with existing Arabic.", "metrics": null }, { "id": "a-new-benchmark-for-evaluating-automatic-speech-recognition-in-the-ara", "name": "A New Benchmark for Evaluating Automatic Speech Recognition in the Arabic Call Domain", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "chat", "asr", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2403.04280" }, "dialects": [ "mixed" ], "year": 2024, "venue": "arXiv 2024", "notes": "This work is an attempt to introduce a comprehensive benchmark for Arabic speech recognition.", "metrics": null }, { "id": "a-survey-of-llms-for-arabic-language-and-its-dialects", "name": "A Survey of LLMs for Arabic Language and its Dialects", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "year": 2024, "venue": "arXiv 2024", "tasks": [ "survey", "llm", "dialects" ], "links": { "paper": "https://arxiv.org/abs/2410.20238" }, "notes": "Survey of LLMs for Arabic and its dialects.", "metrics": null }, { "id": "alclam-arabic-dialectal-language-model", "name": "AlcLaM: Arabic Dialectal Language Model", "type": "paper", "country": "INTL", "org": "amurtadha", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2407.13097", "github": "https://github.com/amurtadha/Alclam", "hf": "https://huggingface.co/rahbi" }, "dialects": [ "msa", "mixed" ], "year": 2024, "venue": "arXiv 2024", "notes": "To tackle this, we construct an Arabic dialectal corpus comprising 3.4M sentences gathered from social media platforms.", "metrics": null }, { "id": "allam-large-language-models-for-arabic-and-english", "name": "ALLaM: Large Language Models for Arabic and English", "type": "paper", "country": "SA", "org": "SDAIA & IBM", "license": "unknown", "modality": "text", "year": 2024, "venue": "arXiv 2024", "tasks": [ "pretraining", "llm" ], "links": { "paper": "https://arxiv.org/abs/2407.15390" }, "notes": "ALLaM: Arabic and English LLMs via second-language acquisition-style training.", "metrics": null }, { "id": "arabiangpt-native-arabic-gpt-based-llm", "name": "ArabianGPT: Native Arabic GPT-based LLM", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "year": 2024, "venue": "arXiv 2024", "tasks": [ "pretraining", "llm" ], "links": { "paper": "https://arxiv.org/abs/2402.15313" }, "notes": "Native Arabic GPT-style LLM family.", "metrics": null }, { "id": "arabic-dataset-for-llm-safeguard-evaluation", "name": "Arabic Dataset for LLM Safeguard Evaluation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2410.17040" }, "year": 2024, "venue": "arXiv 2024", "notes": "In particular, we present an Arab-region-specific safety evaluation dataset consisting of 5,799 questions, including direct attacks, indirect attacks.", "metrics": null }, { "id": "arabic-speech-recognition-of-zero-resourced-languages-a-case-of-shehri", "name": "Arabic Speech Recognition of zero-resourced Languages: A case of Shehri (Jibbali) Language", "type": "paper", "country": "INTL", "org": "College of Computer and Information Scienc", "license": "unknown", "modality": "speech", "tasks": [ "asr" ], "links": { "paper": "https://aclanthology.org/2024.osact-1.10/" }, "year": 2024, "venue": "OSACT 2024", "notes": "We collected a Shehri (Jibbali) speech corpus and utilized transfer learning by fine-tuning pre-trained ASR models on this dataset.", "metrics": null }, { "id": "arabic-text-sentiment-analysis-reinforcing-human-performed-surveys-wit", "name": "Arabic Text Sentiment Analysis: Reinforcing Human-Performed Surveys with Wider Topic Analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://arxiv.org/abs/2403.01921" }, "dialects": [ "msa", "mixed" ], "year": 2024, "venue": "arXiv 2024", "notes": "Sentiment analysis (SA) has been, and is still, a thriving research area.", "metrics": null }, { "id": "arabic-nougat-fine-tuning-vision-transformers-for-arabic-ocr-and-markd", "name": "Arabic-Nougat: Fine-Tuning Vision Transformers for Arabic OCR and Markdown Extraction", "type": "paper", "country": "INTL", "org": "mohamedalirashad", "license": "unknown", "modality": "text", "tasks": [ "ocr", "vision" ], "links": { "paper": "https://arxiv.org/abs/2411.17835", "github": "https://github.com/MohamedAliRashad/arabic-nougat" }, "year": 2024, "venue": "arXiv 2024", "notes": "We present Arabic-Nougat, a suite of OCR models for converting Arabic book pages into structured Markdown text.", "metrics": null }, { "id": "arabicnlu-2024-the-first-arabic-natural-language-understanding-shared", "name": "ArabicNLU 2024: The First Arabic Natural Language Understanding Shared Task", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2407.20663" }, "year": 2024, "venue": "arXiv 2024", "notes": "This paper presents an overview of the Arabic Natural Language Understanding (ArabicNLU 2024) shared task, focusing on two subtasks.", "metrics": null }, { "id": "arablegaleval-a-multitask-benchmark-for-assessing-arabic-legal-knowled", "name": "ArabLegalEval: A Multitask Benchmark for Assessing Arabic Legal Knowledge in Large Language Models", "type": "paper", "country": "INTL", "org": "thiqah", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2408.07983", "github": "https://github.com/Thiqah/ArabLegalEval" }, "dialects": [ "gulf" ], "year": 2024, "venue": "arXiv 2024", "notes": "To address this gap, we introduce ArabLegalEval, a multitask benchmark dataset for assessing the Arabic legal knowledge of LLMs.", "metrics": null }, { "id": "arafinnlp-2024-the-first-arabic-financial-nlp-shared-task", "name": "AraFinNLP 2024: The First Arabic Financial NLP Shared Task", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "chat", "translation" ], "links": { "paper": "https://arxiv.org/abs/2407.09818" }, "dialects": [ "msa", "mixed" ], "year": 2024, "venue": "arXiv 2024", "notes": "The expanding financial markets of the Arab world require sophisticated Arabic NLP tools.", "metrics": null }, { "id": "araieval-shared-task-propagandistic-techniques-detection-in-unimodal-a", "name": "ArAIEval Shared Task: Propagandistic Techniques Detection in Unimodal and Multimodal Arabic Content", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2407.04247", "website": "https://araieval.gitlab.io" }, "year": 2024, "venue": "arXiv 2024", "notes": "We present an overview of the second edition of the ArAIEval shared task, organized as part of the ArabicNLP 2024 conference co-located with ACL 2024.", "metrics": null }, { "id": "arapoembert-a-pretrained-language-model-for-arabic-poetry-analysis", "name": "AraPoemBERT: A Pretrained Language Model for Arabic Poetry Analysis", "type": "paper", "country": "SA", "org": "King Saud University", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "pretraining", "poetry" ], "links": { "paper": "https://arxiv.org/abs/2403.12392", "hf": "https://huggingface.co/faisalq" }, "year": 2024, "venue": "arXiv 2024", "notes": "In this paper, we introduce AraPoemBERT, an Arabic language model pretrained exclusively on Arabic poetry text.", "metrics": null }, { "id": "arastem-a-native-arabic-multiple-choice-question-benchmark-for-evaluat", "name": "AraSTEM: A Native Arabic Multiple Choice Question Benchmark for Evaluating LLMs Knowledge In STEM Subjects", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2501.00559" }, "year": 2024, "venue": "arXiv 2024", "notes": "To address this issue, we introduce AraSTEM, a new Arabic multiple-choice question dataset aimed at evaluating LLMs knowledge in STEM subjects.", "metrics": null }, { "id": "aratar-a-corpus-to-support-the-fine-grained-detection-of-hate-speech-t", "name": "AraTar: A Corpus to Support the Fine-grained Detection of Hate Speech Targets in the Arabic Language", "type": "paper", "country": "INTL", "org": "The University of Manchester", "license": "unknown", "modality": "text", "tasks": [ "hate-speech" ], "links": { "paper": "https://aclanthology.org/2024.osact-1.1/" }, "year": 2024, "venue": "OSACT 2024", "notes": "To comprehensively address this problem, we have combined and re-annotated hate speech tweets from existing publicly available corpora, resulting.", "metrics": null }, { "id": "areeg-chars-dataset-for-envisioned-speech-recognition-using-eeg-for-ar", "name": "ArEEG_Chars: Dataset for Envisioned Speech Recognition using EEG for Arabic Characters", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "speech", "vision" ], "links": { "paper": "https://arxiv.org/abs/2402.15733" }, "year": 2024, "venue": "arXiv 2024", "notes": "We introduce ArEEGChars, a novel EEG dataset for Arabic 31 characters collected from 30 participants.", "metrics": null }, { "id": "areeg-words-dataset-for-envisioned-speech-recognition-using-eeg-for-ar", "name": "ArEEG_Words: Dataset for Envisioned Speech Recognition using EEG for Arabic Words", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "translation", "speech", "vision" ], "links": { "paper": "https://arxiv.org/abs/2411.18888" }, "year": 2024, "venue": "arXiv 2024", "notes": "We introduce in this paper ArEEGWords dataset, a novel EEG dataset recorded from 22 participants with mean age of 22 years.", "metrics": null }, { "id": "arzen-llm-code-switched-egyptian-arabic-english-translation-and-speech", "name": "ArzEn-LLM: Code-Switched Egyptian Arabic-English Translation and Speech Recognition Using LLMs", "type": "paper", "country": "INTL", "org": "ahmedheakl", "license": "unknown", "modality": "speech", "tasks": [ "asr", "translation", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2406.18120", "github": "http://github.com/ahmedheakl/arazn-llm", "hf": "http://huggingface.co/collections/ahmedheakl/arazn-llm-662ceaf12777656607b9524e" }, "dialects": [ "egy" ], "year": 2024, "venue": "arXiv 2024", "notes": "Motivated by the widespread increase in the phenomenon of code-switching between Egyptian Arabic and English in recent times.", "metrics": null }, { "id": "athar-a-high-quality-and-diverse-dataset-for-classical-arabic-to-engli", "name": "ATHAR: A High-Quality and Diverse Dataset for Classical Arabic to English Translation", "type": "paper", "country": "INTL", "org": "mohamed-khalil", "license": "cc-by-sa-4.0", "modality": "text", "tasks": [ "translation", "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2407.19835", "hf": "https://huggingface.co/datasets/mohamed-khalil/ATHAR" }, "dialects": [ "classical" ], "year": 2024, "venue": "arXiv 2024", "notes": "We present the ATHAR dataset, which comprises 66,000 high-quality classical Arabic to English translation samples that cover a wide array of topics.", "metrics": { "downloads": 82, "likes": 13, "lastModified": "2024-08-04" } }, { "id": "automated-essay-scoring-in-arabic-a-dataset-and-analysis-of-a-bert-bas", "name": "Automated essay scoring in Arabic: a dataset and analysis of a BERT-based system", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2407.11212" }, "year": 2024, "venue": "arXiv 2024", "notes": "Automated Essay Scoring (AES) holds significant promise in the field of education, helping educators to mark larger volumes of essays.", "metrics": null }, { "id": "benchmarking-llama-3-on-arabic-language-generation-tasks", "name": "Benchmarking LLaMA-3 on Arabic Language Generation Tasks", "type": "paper", "country": "INTL", "org": "The University of British Col", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://aclanthology.org/2024.arabicnlp-1.24/" }, "dialects": [ "msa" ], "year": 2024, "venue": "ArabicNLP 2024", "notes": "We focus to bridge this gap by evaluating LLaMA-3-70B on a diverse set of Arabic natural language generation (NLG) benchmarks.", "metrics": null }, { "id": "bimedix2-bio-medical-expert-lmm-for-diverse-medical-modalities", "name": "BiMediX2: Bio-Medical EXpert LMM for Diverse Medical Modalities", "type": "paper", "country": "AE", "org": "MBZUAI", "license": "unknown", "modality": "multimodal", "tasks": [ "chat", "qa", "summarization", "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2412.07769", "github": "https://github.com/mbzuai-oryx/BiMediX2" }, "year": 2024, "venue": "arXiv 2024", "notes": "We introduce BiMediX2, a bilingual (Arabic-English) Bio-Medical EXpert Large Multimodal Model that supports text-based and image-based medical interactions.", "metrics": null }, { "id": "bimedix-bilingual-medical-mixture-of-experts-llm", "name": "BiMediX: Bilingual Medical Mixture of Experts LLM", "type": "paper", "country": "AE", "org": "MBZUAI", "license": "unknown", "modality": "text", "tasks": [ "chat", "translation", "qa", "instruction-tuning", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2402.13253", "github": "https://github.com/mbzuai-oryx/BiMediX" }, "year": 2024, "venue": "arXiv 2024", "notes": "In this paper, we introduce BiMediX, the first bilingual medical mixture of experts LLM designed for seamless interaction in both English and Arabic.", "metrics": null }, { "id": "cafe-a-novel-code-switching-dataset-for-algerian-dialect-french-and-en", "name": "CAFE A Novel Code switching Dataset for Algerian Dialect French and English", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "code-switching", "asr", "speech" ], "links": { "paper": "https://arxiv.org/abs/2411.13424" }, "dialects": [ "magh" ], "year": 2024, "venue": "arXiv 2024", "notes": "CAFE: first code-switching speech dataset mixing Algerian dialect, French and English, from spontaneous human-human conversations.", "metrics": null }, { "id": "camel-bench-a-comprehensive-arabic-lmm-benchmark", "name": "CAMEL-Bench: A Comprehensive Arabic LMM Benchmark", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "ocr", "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2410.18976" }, "year": 2024, "venue": "arXiv 2024", "notes": "Recent years have witnessed a significant interest in developing large multimodal models.", "metrics": null }, { "id": "cameleval-advancing-culturally-aligned-arabic-language-models-and-benc", "name": "CamelEval: Advancing Culturally Aligned Arabic Language Models and Benchmarks", "type": "paper", "country": "INTL", "org": "elmrc", "license": "unknown", "modality": "text", "tasks": [ "qa", "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2409.12623", "hf": "https://huggingface.co/elmrc" }, "year": 2024, "venue": "arXiv 2024", "notes": "This paper introduces Juhaina, a Arabic-English bilingual LLM specifically designed to align with the values and preferences of Arabic speakers.", "metrics": null }, { "id": "catt-character-based-arabic-tashkeel-transformer", "name": "CATT: Character-based Arabic Tashkeel Transformer", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "year": 2024, "venue": "arXiv 2024", "tasks": [ "diacritization" ], "links": { "paper": "https://arxiv.org/abs/2407.03236", "github": "https://github.com/abjadai/catt" }, "notes": "Character-based Transformer for Arabic diacritization (tashkeel).", "metrics": null }, { "id": "cidar-culturally-relevant-instruction-dataset-for-arabic", "name": "CIDAR: Culturally Relevant Instruction Dataset For Arabic", "type": "paper", "country": "SA", "org": "ARBML", "license": "unknown", "modality": "text", "tasks": [ "instruction-tuning" ], "links": { "paper": "https://arxiv.org/abs/2402.03177", "github": "https://github.com/ARBML/CIDAR" }, "dialects": [ "lev" ], "year": 2024, "venue": "arXiv 2024", "notes": "Instruction tuning has emerged as a prominent methodology for teaching Large Language Models (LLMs) to follow instructions.", "metrics": null }, { "id": "cross-attention-fusion-of-visual-and-geometric-features-for-large-voca", "name": "Cross-Attention Fusion of Visual and Geometric Features for Large Vocabulary Arabic Lipreading", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "speech", "vision" ], "links": { "paper": "https://arxiv.org/abs/2402.11520", "website": "https://crns-smartvision.github.io/lrwar" }, "year": 2024, "venue": "arXiv 2024", "notes": "To address this challenge, firstly, we propose a cross-attention fusion-based approach for large lexicon Arabic vocabulary to predict spoken words in videos.", "metrics": null }, { "id": "darijabanking-a-new-resource-for-overcoming-language-barriers-in-banki", "name": "DarijaBanking: A New Resource for Overcoming Language Barriers in Banking Intent Detection for Moroccan Arabic Speakers", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "chat", "embedding" ], "links": { "paper": "https://arxiv.org/abs/2405.16482" }, "dialects": [ "magh", "msa" ], "year": 2024, "venue": "arXiv 2024", "notes": "Navigating the complexities of language diversity is a central challenge in developing robust natural language processing systems.", "metrics": null }, { "id": "dialectal-pretraining-for-arabic-asr-mbzuai", "name": "Dialectal pretraining for Arabic ASR (MBZUAI)", "type": "paper", "country": "AE", "org": "MBZUAI", "license": "unknown", "modality": "speech", "tasks": [ "asr" ], "links": { "paper": "https://arxiv.org/abs/2411.05872" }, "year": 2024, "venue": "arXiv 2024", "notes": "Study and models showing dialectal pretraining improves Arabic automatic speech recognition.", "metrics": null }, { "id": "egybert-bert-pretrained-on-egyptian-dialect-corpora", "name": "EgyBERT: BERT Pretrained on Egyptian Dialect Corpora", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "year": 2024, "venue": "arXiv 2024", "tasks": [ "pretraining", "encoder", "dialects" ], "links": { "paper": "https://arxiv.org/abs/2408.03524" }, "notes": "BERT pretrained on Egyptian dialect corpora.", "metrics": null }, { "id": "estimating-the-level-of-dialectness-predicts-interannotator-agreement", "name": "Estimating the Level of Dialectness Predicts Interannotator Agreement in Multi-dialect Arabic Datasets", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id" ], "links": { "paper": "https://arxiv.org/abs/2405.11282" }, "dialects": [ "mixed" ], "year": 2024, "venue": "arXiv 2024", "notes": "On annotating multi-dialect Arabic datasets, it is common to randomly assign the samples across a pool of native Arabic speakers.", "metrics": null }, { "id": "event-arguments-extraction-corpus-and-modeling-using-bert-for-arabic", "name": "Event-Arguments Extraction Corpus and Modeling using BERT for Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2407.21153", "website": "https://sina.birzeit.edu/wojood" }, "year": 2024, "venue": "arXiv 2024", "notes": "To fill this gap, we introduce the \\hadath corpus ($550$k tokens) as an extension of Wojood, enriched with event-argument annotations.", "metrics": null }, { "id": "exploiting-dialect-identification-in-automatic-dialectal-text-normaliz", "name": "Exploiting Dialect Identification in Automatic Dialectal Text Normalization", "type": "paper", "country": "AE", "org": "CAMeL Lab, NYUAD", "license": "unknown", "modality": "text", "tasks": [ "dialect-id" ], "links": { "paper": "https://aclanthology.org/2024.arabicnlp-1.4/" }, "year": 2024, "venue": "ArabicNLP 2024", "notes": "We benchmark newly developed pretrained sequence-to-sequence models on the task of CODAfication.", "metrics": null }, { "id": "fineweb-edu-ar-machine-translated-corpus-to-support-arabic-small-langu", "name": "Fineweb-Edu-Ar: Machine-translated Corpus to Support Arabic Small Language Models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation" ], "links": { "paper": "https://arxiv.org/abs/2411.06402" }, "year": 2024, "venue": "arXiv 2024", "notes": "This is especially true for multilingual LLMs, where the scarcity of high-quality and readily available data online has led to a multitude.", "metrics": null }, { "id": "from-nile-sands-to-digital-hands-machine-translation-of-coptic-texts", "name": "From Nile Sands to Digital Hands: Machine Translation of Coptic Texts", "type": "paper", "country": "AE", "org": "MBZUAI", "license": "unknown", "modality": "text", "tasks": [ "translation" ], "links": { "paper": "https://aclanthology.org/2024.arabicnlp-1.25/", "github": "https://github.com/UBC-NLP/copticmt" }, "year": 2024, "venue": "ArabicNLP 2024", "notes": "We have also developed the first neural machine translation system between Coptic, English, and Arabic.", "metrics": null }, { "id": "gazelle-an-instruction-dataset-for-arabic-writing-assistance", "name": "Gazelle: An Instruction Dataset for Arabic Writing Assistance", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "instruction-tuning", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2410.18163" }, "year": 2024, "venue": "arXiv 2024", "notes": "To address these issues, we present Gazelle, a comprehensive dataset for Arabic writing assistance.", "metrics": null }, { "id": "glare-google-apps-arabic-reviews-dataset", "name": "GLARE: Google Apps Arabic Reviews Dataset", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "nlp" ], "links": { "paper": "https://arxiv.org/abs/2412.15259" }, "dialects": [ "gulf" ], "year": 2024, "venue": "arXiv 2024", "notes": "This paper introduces GLARE an Arabic Apps Reviews dataset collected from Saudi Google PlayStore.", "metrics": null }, { "id": "hate-speech-detection-in-arabic-corpus-design-and-evaluation", "name": "Hate speech detection in Arabic: corpus design and evaluation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "year": 2024, "venue": "Frontiers in AI 2024", "tasks": [ "hate-speech", "dataset" ], "links": { "paper": "https://www.frontiersin.org/journals/artificial-intelligence/articles/10.3389/frai.2024.1345445/full" }, "notes": "Arabic hate speech detection: corpus design and evaluation.", "metrics": null }, { "id": "humvi-a-multilingual-dataset-for-detecting-violent-incidents-impacting", "name": "HumVI: A Multilingual Dataset for Detecting Violent Incidents Impacting Humanitarian Aid", "type": "paper", "country": "INTL", "org": "dataminr-ai", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2410.06370", "github": "https://github.com/dataminr-ai/humvi-dataset" }, "year": 2024, "venue": "arXiv 2024", "notes": "HumVI: English, French and Arabic news articles labelled for violent incidents affecting humanitarian aid.", "metrics": null }, { "id": "leveraging-corpus-metadata-to-detect-template-based-translation-an-exp", "name": "Leveraging Corpus Metadata to Detect Template-based Translation: An Exploratory Case Study of the Egyptian Arabic Wikipedia Edition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation" ], "links": { "paper": "https://arxiv.org/abs/2404.00565" }, "dialects": [ "egy", "magh" ], "year": 2024, "venue": "arXiv 2024", "notes": "Wikipedia articles (content pages) are commonly used corpora in Natural Language Processing.", "metrics": null }, { "id": "llamalens-specialized-multilingual-llm-for-analyzing-news-and-social-m", "name": "LlamaLens: Specialized Multilingual LLM for Analyzing News and Social Media Content", "type": "paper", "country": "QA", "org": "QCRI", "license": "unknown", "modality": "text", "tasks": [ "news", "social-media", "multilingual" ], "links": { "paper": "https://arxiv.org/abs/2410.15308", "hf": "https://huggingface.co/collections/QCRI/llamalens-672f7e0604a0498c6a2f0fe9" }, "year": 2024, "venue": "arXiv 2024", "notes": "LlamaLens: specialised multilingual LLM for news and social media analysis, tested on 18 tasks and 52 Arabic, English and Hindi datasets.", "metrics": null }, { "id": "llm-based-mt-data-creation-dialectal-to-msa-translation-shared-task", "name": "LLM-based MT Data Creation: Dialectal to MSA Translation Shared Task", "type": "paper", "country": "INTL", "org": "aiXplain Inc.", "license": "unknown", "modality": "text", "tasks": [ "translation", "shared-task" ], "links": { "paper": "https://aclanthology.org/2024.osact-1.14/" }, "dialects": [ "msa" ], "year": 2024, "venue": "OSACT 2024", "notes": "This paper presents our approach to the Dialect to Modern Standard Arabic (MSA) Machine Translation shared task, conducted as part of the sixth Workshop.", "metrics": null }, { "id": "mememind-at-araieval-shared-task-spotting-persuasive-spans-in-arabic-t", "name": "MemeMind at ArAIEval Shared Task: Spotting Persuasive Spans in Arabic Text with Persuasion Techniques Identification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2408.04540" }, "year": 2024, "venue": "arXiv 2024", "notes": "Using attention masks, we created uniform lengths for each span and assigned BIO tags to each token based on the provided labels.", "metrics": null }, { "id": "mentalqa-an-annotated-arabic-corpus-for-questions-and-answers-of-menta", "name": "MentalQA: An Annotated Arabic Corpus for Questions and Answers of Mental Healthcare", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "chat", "sentiment", "qa", "vision" ], "links": { "paper": "https://arxiv.org/abs/2405.12619" }, "year": 2024, "venue": "arXiv 2024", "notes": "We introduce MentalQA, a novel Arabic dataset featuring conversational-style question-and-answer (QA) interactions.", "metrics": null }, { "id": "muharaf-manuscripts-of-handwritten-arabic-dataset-for-cursive-text-rec", "name": "Muharaf: Manuscripts of Handwritten Arabic Dataset for Cursive Text Recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "asr", "ocr", "vision", "poetry" ], "links": { "paper": "https://arxiv.org/abs/2406.09630" }, "year": 2024, "venue": "arXiv 2024", "notes": "We present the Manuscripts of Handwritten Arabic(Muharaf) dataset.", "metrics": null }, { "id": "nadi-2024-fifth-nuanced-arabic-dialect-identification", "name": "NADI 2024: Fifth Nuanced Arabic Dialect Identification Shared Task", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "year": 2024, "venue": "ArabicNLP 2024", "tasks": [ "dialect-id", "shared-task" ], "links": { "paper": "https://arxiv.org/abs/2407.04910" }, "notes": "Overview of the fifth Nuanced Arabic Dialect Identification shared task.", "metrics": null }, { "id": "natural-language-processing-for-dialects-of-a-language-a-survey", "name": "Natural Language Processing for Dialects of a Language: A Survey", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "dialect-id", "sentiment", "summarization", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2401.05632" }, "dialects": [ "mixed" ], "year": 2024, "venue": "arXiv 2024", "notes": "Survey of NLP for dialects of a language, covering datasets and approaches for dialectal NLU and NLG, Arabic dialects included.", "metrics": null }, { "id": "nullpointer-at-araieval-shared-task-arabic-propagandist-technique-dete", "name": "Nullpointer at ArAIEval Shared Task: Arabic Propagandist Technique Detection with Token-to-Word Mapping in Sequence Tagging", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2407.01360" }, "year": 2024, "venue": "arXiv 2024", "notes": "This paper investigates the optimization of propaganda technique detection in Arabic text, including tweets \\& news paragraphs, from ArAIEval shared task 1.", "metrics": null }, { "id": "on-the-importance-of-data-scale-in-pretraining-arabic-language-models", "name": "On the importance of Data Scale in Pretraining Arabic Language Models", "type": "paper", "country": "INTL", "org": "huawei-noah", "license": "unknown", "modality": "text", "tasks": [ "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2401.07760", "github": "https://github.com/huawei-noah/Pretrained-Language-Model/tree/master/JABER-PyTorch" }, "year": 2024, "venue": "arXiv 2024", "notes": "Pretraining monolingual language models have been proven to be vital for performance in Arabic Natural Language Processing (NLP) tasks.", "metrics": null }, { "id": "optimized-quran-passage-retrieval-using-an-expanded-qa-dataset-and-fin", "name": "Optimized Quran Passage Retrieval Using an Expanded QA Dataset and Fine-Tuned Language Models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "qa", "quran" ], "links": { "paper": "https://arxiv.org/abs/2412.11431" }, "dialects": [ "classical", "msa" ], "year": 2024, "venue": "arXiv 2024", "notes": "Understanding the deep meanings of the Qur'an and bridging the language gap between modern standard Arabic and classical Arabic is essential.", "metrics": null }, { "id": "osact6-dialect-to-msa-translation-shared-task-overview", "name": "OSACT6 Dialect to MSA Translation Shared Task Overview", "type": "paper", "country": "INTL", "org": "aiXplain Inc.", "license": "unknown", "modality": "text", "tasks": [ "translation", "shared-task" ], "links": { "paper": "https://aclanthology.org/2024.osact-1.11/" }, "dialects": [ "msa" ], "year": 2024, "venue": "OSACT 2024", "notes": "This paper presents the Dialectal Arabic (DA) to Modern Standard Arabic (MSA) Machine Translation (MT) shared task in the sixth Workshop on Open-Source.", "metrics": null }, { "id": "palo-a-polyglot-large-multimodal-model-for-5b-people", "name": "PALO: A Polyglot Large Multimodal Model for 5B People", "type": "paper", "country": "AE", "org": "MBZUAI", "license": "unknown", "modality": "multimodal", "tasks": [ "translation", "instruction-tuning", "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2402.14818", "github": "https://github.com/mbzuai-oryx/PALO" }, "year": 2024, "venue": "arXiv 2024", "notes": "In pursuit of more inclusive Vision-Language Models (VLMs), this study introduces a Large Multilingual Multimodal Model called PALO.", "metrics": null }, { "id": "peacock-a-family-of-arabic-multimodal-large-language-models-and-benchm", "name": "Peacock: A Family of Arabic Multimodal Large Language Models and Benchmarks", "type": "paper", "country": "INTL", "org": "UBC-NLP", "license": "unknown", "modality": "multimodal", "tasks": [ "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2403.01031", "github": "https://github.com/UBC-NLP/peacock" }, "year": 2024, "venue": "arXiv 2024", "notes": "Multimodal large language models (MLLMs) have proven effective in a wide range of tasks requiring complex reasoning and linguistic comprehension.", "metrics": null }, { "id": "performance-analysis-of-speech-encoders-for-low-resource-slu-and-asr-i", "name": "Performance Analysis of Speech Encoders for Low-Resource SLU and ASR in Tunisian Dialect", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "asr", "pretraining", "speech" ], "links": { "paper": "https://arxiv.org/abs/2407.04533" }, "dialects": [ "magh" ], "year": 2024, "venue": "arXiv 2024", "notes": "Speech encoders pretrained through self-supervised learning (SSL) have demonstrated remarkable performance in various downstream tasks.", "metrics": null }, { "id": "propaganda-to-hate-a-multimodal-analysis-of-arabic-memes-with-multi-ag", "name": "Propaganda to Hate: A Multimodal Analysis of Arabic Memes with Multi-Agent LLMs", "type": "paper", "country": "INTL", "org": "firojalam", "license": "unknown", "modality": "multimodal", "tasks": [ "vision" ], "links": { "paper": "https://arxiv.org/abs/2409.07246", "github": "https://github.com/firojalam/propaganda-and-hateful-memes" }, "year": 2024, "venue": "arXiv 2024", "notes": "In the past decade, social media platforms have been used for information dissemination and consumption.", "metrics": null }, { "id": "qabas-an-open-source-arabic-lexicographic-database", "name": "Qabas: An Open-Source Arabic Lexicographic Database", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "nlp" ], "links": { "paper": "https://arxiv.org/abs/2406.06598", "website": "https://sina.birzeit.edu/qabas" }, "year": 2024, "venue": "arXiv 2024", "notes": "We present Qabas, a novel open-source Arabic lexicon designed for NLP applications.", "metrics": null }, { "id": "quranic-audio-dataset-crowdsourced-and-labeled-recitation-from-non-ara", "name": "Quranic Audio Dataset: Crowdsourced and Labeled Recitation from Non-Arabic Speakers", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "speech", "quran" ], "links": { "paper": "https://arxiv.org/abs/2405.02675" }, "dialects": [ "classical" ], "year": 2024, "venue": "arXiv 2024", "notes": "This paper addresses the challenge of learning to recite the Quran for non-Arabic speakers.", "metrics": null }, { "id": "receiptsense-beyond-traditional-ocr-a-dataset-for-receipt-understandin", "name": "ReceiptSense: Beyond Traditional OCR -- A Dataset for Receipt Understanding", "type": "paper", "country": "INTL", "org": "update-for-integrated-business-ai", "license": "unknown", "modality": "vision", "tasks": [ "ocr", "qa", "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2406.04493", "github": "https://github.com/Update-For-Integrated-Business-AI/CORU" }, "year": 2024, "venue": "arXiv 2024", "notes": "We introduce \\dataset, a comprehensive dataset designed for Arabic-English receipt understanding comprising 20,000 annotated receipts.", "metrics": null }, { "id": "resource-aware-arabic-llm-creation-model-adaptation-integration-and-mu", "name": "Resource-Aware Arabic LLM Creation: Model Adaptation, Integration, and Multi-Domain Testing", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "diacritization", "dialect-id", "qa", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2412.17548" }, "year": 2024, "venue": "arXiv 2024", "notes": "This paper presents a novel approach to fine-tuning the Qwen2-1.5B model for Arabic language processing using Quantized Low-Rank Adaptation.", "metrics": null }, { "id": "sambalingo-teaching-llms-new-languages", "name": "SambaLingo: Teaching LLMs New Languages", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "year": 2024, "venue": "arXiv 2024", "tasks": [ "pretraining", "multilingual", "llm" ], "links": { "paper": "https://arxiv.org/abs/2404.05829" }, "notes": "Recipe for adapting LLMs to new languages including Arabic.", "metrics": null }, { "id": "saudibert-bert-pretrained-on-saudi-dialect-corpora", "name": "SaudiBERT: BERT Pretrained on Saudi Dialect Corpora", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "year": 2024, "venue": "arXiv 2024", "tasks": [ "pretraining", "encoder", "dialects" ], "links": { "paper": "https://arxiv.org/abs/2405.06239" }, "notes": "BERT pretrained on Saudi dialect corpora.", "metrics": null }, { "id": "sina-at-fignews-2024-multilingual-datasets-annotated-with-bias-and-pro", "name": "Sina at FigNews 2024: Multilingual Datasets Annotated with Bias and Propaganda", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2407.09327", "website": "https://sina.birzeit.edu/fada" }, "year": 2024, "venue": "arXiv 2024", "notes": "The proliferation of bias and propaganda on social media is an increasingly significant concern.", "metrics": null }, { "id": "sinatools-open-source-toolkit-for-arabic-natural-language-processing", "name": "SinaTools: Open Source Toolkit for Arabic Natural Language Processing", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "diacritization", "ner", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2411.01523", "website": "https://sina.birzeit.edu/sinatools" }, "year": 2024, "venue": "arXiv 2024", "notes": "We introduce SinaTools, an open-source Python package for Arabic natural language processing and understanding.", "metrics": null }, { "id": "strategies-for-arabic-readability-modeling", "name": "Strategies for Arabic Readability Modeling", "type": "paper", "country": "AE", "org": "CAMeL Lab, NYUAD", "license": "unknown", "modality": "text", "tasks": [ "readability" ], "links": { "paper": "https://aclanthology.org/2024.arabicnlp-1.5/" }, "year": 2024, "venue": "ArabicNLP 2024", "notes": "We present a set of experimental results on Arabic readability assessment using a diverse range of approaches, from rule-based methods to Arabic pretrained.", "metrics": null }, { "id": "swan-and-arabicmteb-dialect-aware-cross-lingual-language", "name": "Swan and ArabicMTEB: Dialect-Aware, Cross-Lingual Language Understanding", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "year": 2024, "venue": "arXiv 2024", "tasks": [ "embedding", "benchmark" ], "links": { "paper": "https://arxiv.org/abs/2411.01192" }, "notes": "Dialect-aware Swan embedding models and the ArabicMTEB benchmark.", "metrics": null }, { "id": "synthetic-arabic-medical-dialogues-using-advanced-multi-agent-llm-tech", "name": "Synthetic Arabic Medical Dialogues Using Advanced Multi-Agent LLM Techniques", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "medical", "dialogue" ], "links": { "paper": "https://aclanthology.org/2024.arabicnlp-1.2/" }, "year": 2024, "venue": "ArabicNLP 2024", "notes": "We introduce a novel Multi-Agent LLM approach capable of generating synthetic Arabic medical dialogues from patient notes, regardless of the original language.", "metrics": null }, { "id": "tafsirextractor-text-preprocessing-pipeline-preparing-classical-arabic", "name": "TafsirExtractor: Text Preprocessing Pipeline preparing Classical Arabic Literature for Machine Learning Applications", "type": "paper", "country": "INTL", "org": "Goethe University Frankfurt", "license": "unknown", "modality": "text", "tasks": [ "ner" ], "links": { "paper": "https://aclanthology.org/2024.osact-1.8/" }, "dialects": [ "classical" ], "year": 2024, "venue": "OSACT 2024", "notes": "We present a comprehensive tool of preprocessing Classical Arabic (CA) literature in the field of historical exegetical studies for machine learning (ML)", "metrics": null }, { "id": "the-evolution-of-darija-open-dataset-introducing-version-2", "name": "The Evolution of Darija Open Dataset: Introducing Version 2", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2405.13016" }, "dialects": [ "magh", "mixed" ], "year": 2024, "venue": "arXiv 2024", "notes": "Darija Open Dataset (DODa) represents an open-source project aimed at enhancing Natural Language Processing capabilities for the Moroccan dialect, Darija.", "metrics": null }, { "id": "the-fignews-shared-task-on-news-media-narratives", "name": "The FIGNEWS Shared Task on News Media Narratives", "type": "paper", "country": "QA", "org": "Northwestern University in Qatar", "license": "unknown", "modality": "text", "tasks": [ "shared-task" ], "links": { "paper": "https://aclanthology.org/2024.arabicnlp-1.56/" }, "year": 2024, "venue": "ArabicNLP 2024", "notes": "We present an overview of the FIGNEWSshared task, organized as part of the Arabic-NLP 2024 conference co-located with ACL2024.", "metrics": null }, { "id": "the-qiyas-benchmark-measuring-chatgpt-mathematical-and-language-unders", "name": "The Qiyas Benchmark: Measuring ChatGPT Mathematical and Language Understanding in Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "reasoning", "math", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2407.00146" }, "year": 2024, "venue": "arXiv 2024", "notes": "Qiyas benchmarks: Arabic mathematical reasoning and language understanding questions from Saudi's GAT exam, tested on ChatGPT.", "metrics": null }, { "id": "the-samer-arabic-text-simplification-corpus", "name": "The SAMER Arabic Text Simplification Corpus", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "nlp" ], "links": { "paper": "https://arxiv.org/abs/2404.18615" }, "year": 2024, "venue": "arXiv 2024", "notes": "We present the SAMER Corpus, the first manually annotated Arabic parallel corpus for text simplification targeting school-aged learners.", "metrics": null }, { "id": "the-translation-of-circumlocution-in-arabic-short-stories-into-english", "name": "The Translation of Circumlocution in Arabic Short Stories into English", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation" ], "links": { "paper": "https://arxiv.org/abs/2411.02887", "website": "https://ntu.edu.iq" }, "year": 2024, "venue": "arXiv 2024", "notes": "This study investigates the translation of circumlocution from Arabic to English in a corpus of short stories by renowned Arabic authors.", "metrics": null }, { "id": "tibyan-corpus-balanced-and-comprehensive-error-coverage-corpus-using-c", "name": "Tibyan Corpus: Balanced and Comprehensive Error Coverage Corpus Using ChatGPT for Arabic Grammatical Error Correction", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "chat" ], "links": { "paper": "https://arxiv.org/abs/2411.04588" }, "year": 2024, "venue": "arXiv 2024", "notes": "Natural language processing (NLP) utilizes text data augmentation to overcome sample size constraints.", "metrics": null }, { "id": "to-distill-or-not-to-distill-on-the-robustness-of-robust-knowledge-dis", "name": "To Distill or Not to Distill? On the Robustness of Robust Knowledge Distillation", "type": "paper", "country": "INTL", "org": "UBC-NLP", "license": "unknown", "modality": "speech", "tasks": [ "asr", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2406.04512" }, "dialects": [ "mixed" ], "year": 2024, "venue": "arXiv 2024", "notes": "Arabic is known to present unique challenges for Automatic Speech Recognition (ASR).", "metrics": null }, { "id": "towards-global-ai-inclusivity-a-large-scale-multilingual-terminology-d", "name": "Towards Global AI Inclusivity: A Large-Scale Multilingual Terminology Dataset (GIST)", "type": "paper", "country": "INTL", "org": "jerry999", "license": "apache-2.0", "modality": "text", "tasks": [ "translation", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2412.18367", "hf": "https://huggingface.co/datasets/Jerry999/multilingual-terminology" }, "year": 2024, "venue": "arXiv 2024", "notes": "We introduce GIST, a large-scale multilingual AI terminology dataset containing 5K terms extracted from top AI conference papers spanning 2000 to 2023.", "metrics": { "downloads": 72, "likes": 1, "lastModified": "2025-05-31" } }, { "id": "towards-zero-shot-text-to-speech-for-arabic-dialects", "name": "Towards Zero-Shot Text-To-Speech for Arabic Dialects", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "tts", "dialect-id", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2406.16751", "website": "https://docs.coqui.ai/en/latest/models/xtts.html" }, "dialects": [ "mixed" ], "year": 2024, "venue": "arXiv 2024", "notes": "Zero-shot multi-speaker text-to-speech (ZS-TTS) systems have advanced for English, however, it still lags behind due to insufficient resources.", "metrics": null }, { "id": "visper-multilingual-audio-visual-speech-recognition", "name": "ViSpeR: Multilingual Audio-Visual Speech Recognition", "type": "paper", "country": "INTL", "org": "yasserdahouml", "license": "unknown", "modality": "speech", "tasks": [ "asr", "evaluation", "speech", "vision" ], "links": { "paper": "https://arxiv.org/abs/2406.00038", "github": "https://github.com/YasserdahouML/visper" }, "year": 2024, "venue": "arXiv 2024", "notes": "This work presents an extensive and detailed study on Audio-Visual Speech Recognition (AVSR) for five widely spoken languages.", "metrics": null }, { "id": "wojoodner-2024-the-second-arabic-named-entity-recognition-shared-task", "name": "WojoodNER 2024: The Second Arabic Named Entity Recognition Shared Task", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "ner" ], "links": { "paper": "https://arxiv.org/abs/2407.09936" }, "year": 2024, "venue": "arXiv 2024", "notes": "We present WojoodNER-2024, the second Arabic Named Entity Recognition (NER) Shared Task.", "metrics": null }, { "id": "xmodel-1-5-an-1b-scale-multilingual-llm", "name": "Xmodel-1.5: An 1B-scale Multilingual LLM", "type": "paper", "country": "INTL", "org": "xiaoduoailab", "license": "unknown", "modality": "text", "tasks": [ "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2411.10083", "github": "https://github.com/XiaoduoAILab/XmodelLM-1.5" }, "year": 2024, "venue": "arXiv 2024", "notes": "We introduce Xmodel-1.5, a 1-billion-parameter multilingual large language model pretrained on 2 trillion tokens.", "metrics": null }, { "id": "xtrust-on-the-multilingual-trustworthiness-of-large-language-models", "name": "XTRUST: On the Multilingual Trustworthiness of Large Language Models", "type": "paper", "country": "INTL", "org": "lluckyyh", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2409.15762", "github": "https://github.com/LluckyYH/XTRUST" }, "year": 2024, "venue": "arXiv 2024", "notes": "In response to the growing global deployment of LLMs, we introduce XTRUST, the first comprehensive multilingual trustworthiness benchmark.", "metrics": null }, { "id": "zaebuc-spoken-a-multilingual-multidialectal-arabic-english-speech-corp", "name": "ZAEBUC-Spoken: A Multilingual Multidialectal Arabic-English Speech Corpus", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "chat", "asr", "speech" ], "links": { "paper": "https://arxiv.org/abs/2403.18182" }, "dialects": [ "egy", "gulf", "msa", "mixed" ], "year": 2024, "venue": "arXiv 2024", "notes": "We present ZAEBUC-Spoken, a multilingual multidialectal Arabic-English speech corpus.", "metrics": null }, { "id": "gptaraeval-a-comprehensive-evaluation-of-chatgpt-on-arabic-n", "name": "GPTAraEval: A Comprehensive Evaluation of ChatGPT on Arabic NLP", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2305.14976" }, "year": 2023, "venue": "Conference on Empirical Methods in Natural Language Processing", "citations": 108, "notes": "ChatGPT's emergence heralds a transformative phase in NLP, particularly demonstrated through its excellent performance on many English benchmarks.", "metrics": null }, { "id": "isolated-arabic-sign-language-recognition-using-a-transforme", "name": "Isolated Arabic Sign Language Recognition Using a Transformer-based Model and Landmark Keypoints", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "multimodal" ], "links": { "paper": "https://doi.org/10.1145/3584984" }, "year": 2023, "venue": "ACM Trans. Asian Low Resour. Lang. Inf. Process.", "citations": 95, "notes": "This article presents a framework for isolated Arabic sign language recognition using hand and face keypoints.", "metrics": null }, { "id": "a-survey-of-ocr-in-arabic-language-applications-techniques-a", "name": "A Survey of OCR in Arabic Language: Applications, Techniques, and Challenges", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "survey", "ocr" ], "links": { "paper": "https://doi.org/10.3390/app13074584" }, "year": 2023, "venue": "Applied Sciences", "citations": 77, "notes": "Optical character recognition (OCR) is the process of extracting handwritten or printed text from a scanned or printed image and converting it to a", "metrics": null }, { "id": "arabic-abstractive-text-summarization-using-rnn-based-and-tr", "name": "Arabic abstractive text summarization using RNN-based and transformer-based architectures", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "summarization" ], "links": { "paper": "https://doi.org/10.1016/j.ipm.2022.103227" }, "year": 2023, "venue": "Information Processing & Management", "citations": 75, "metrics": null }, { "id": "arabic-tweets-based-sentiment-analysis-to-investigate-the-im", "name": "Arabic Tweets-Based Sentiment Analysis to Investigate the Impact of COVID-19 in KSA: A Deep Learning Approach", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.3390/bdcc7010016" }, "year": 2023, "venue": "Big Data and Cognitive Computing", "citations": 70, "notes": "The World Health Organization (WHO) declared the outbreak of Coronavirus disease 2019 (COVID-19) a pandemic on 11 March 2020.", "metrics": null }, { "id": "prediction-of-the-customers-interests-using-sentiment-analys", "name": "Prediction of the customers' interests using sentiment analysis in e-commerce data for comparison of Arabic, English, and Turkish languages", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1016/j.jksuci.2023.02.017" }, "year": 2023, "venue": "Journal of King Saud University: Computer and Information Sciences", "citations": 67, "notes": "In the business world, large companies that can achieve continuity in innovation gain a significant competitive advantage.", "metrics": null }, { "id": "arabic-sentiment-analysis-of-youtube-comments-nlp-based-mach", "name": "Arabic Sentiment Analysis of YouTube Comments: NLP-Based Machine Learning Approaches for Content Evaluation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "evaluation" ], "links": { "paper": "https://doi.org/10.3390/bdcc7030127" }, "year": 2023, "venue": "Big Data and Cognitive Computing", "citations": 61, "notes": "YouTube is a popular video-sharing platform that offers a diverse range of content.", "metrics": null }, { "id": "comparative-performance-of-ensemble-machine-learning-for-ara", "name": "Comparative performance of ensemble machine learning for Arabic cyberbullying and offensive language detection", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification", "evaluation" ], "links": { "paper": "https://doi.org/10.1007/s10579-023-09683-y" }, "year": 2023, "venue": "Language Resources and Evaluation", "citations": 53, "notes": "Since cyberbullying impacts both individual victims and entire society, research on abusive language and its detection has attracted attention in recent years.", "metrics": null }, { "id": "tarjamat-evaluation-of-bard-and-chatgpt-on-machine-translati", "name": "TARJAMAT: Evaluation of Bard and ChatGPT on Machine Translation of Ten Arabic Varieties", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2308.03051" }, "year": 2023, "venue": "ARABICNLP", "citations": 52, "notes": "Considering this constraint, we present a thorough assessment of Bard and ChatGPT", "metrics": null }, { "id": "beyond-english-evaluating-llms-for-arabic-grammatical-error", "name": "Beyond English: Evaluating LLMs for Arabic Grammatical Error Correction", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2312.08400" }, "year": 2023, "venue": "ARABICNLP", "citations": 35, "notes": "Large language models (LLMs) finetuned to follow human instruction have recently exhibited significant capabilities in various English NLP tasks.", "metrics": null }, { "id": "qur-an-qa-2023-shared-task-overview-of-passage-retrieval-and", "name": "Qur’an QA 2023 Shared Task: Overview of Passage Retrieval and Reading Comprehension Tasks over the Holy Qur’an", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa", "quran", "evaluation" ], "links": { "paper": "https://aclanthology.org/2023.arabicnlp-1.76/" }, "year": 2023, "venue": "ARABICNLP", "citations": 31, "notes": "Motivated by the need for intelligent question answering (QA) systems on the Holy Qur’an and the success of the first Qur’an Question Answering shared task", "metrics": null }, { "id": "arabic-dialect-identification-under-scrutiny-limitations-of", "name": "Arabic Dialect Identification under Scrutiny: Limitations of Single-label Classification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id", "classification" ], "links": { "paper": "https://arxiv.org/abs/2310.13661" }, "year": 2023, "venue": "ARABICNLP", "citations": 28, "notes": "Automatic Arabic Dialect Identification (ADI) of text has gained great popularity since it was introduced in the early 2010s.", "metrics": null }, { "id": "a-parameter-efficient-learning-approach-to-arabic-dialect-id", "name": "A Parameter-Efficient Learning Approach to Arabic Dialect Identification with Pre-Trained General-Purpose Speech Model", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id", "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2305.11244" }, "year": 2023, "venue": "Interspeech", "citations": 26, "notes": "In this work, we explore Parameter-Efficient-Learning (PEL) techniques to repurpose a General-Purpose-Speech (GSM) model for Arabic dialect identification", "metrics": null }, { "id": "artst-arabic-text-and-speech-transformer", "name": "ArTST: Arabic Text and Speech Transformer", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "tts", "dialect-id" ], "links": { "paper": "https://arxiv.org/abs/2310.16621" }, "year": 2023, "venue": "ARABICNLP", "citations": 26, "notes": "We present ArTST, a pre-trained Arabic text and speech transformer for supporting open-source speech technologies for the Arabic language.", "metrics": null }, { "id": "evaluating-chatgpt-and-bard-ai-on-arabic-sentiment-analysis", "name": "Evaluating ChatGPT and Bard AI on Arabic Sentiment Analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "evaluation" ], "links": { "paper": "https://aclanthology.org/2023.arabicnlp-1.27/" }, "year": 2023, "venue": "ARABICNLP", "citations": 24, "notes": "Large Language Models (LLMs) such as ChatGPT and Bard AI have gained much attention due to their outstanding performance on a range of NLP tasks.", "metrics": null }, { "id": "octopus-a-multitask-model-and-toolkit-for-arabic-natural-lan", "name": "Octopus: A Multitask Model and Toolkit for Arabic Natural Language Generation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2310.16127" }, "year": 2023, "venue": "ARABICNLP", "citations": 24, "notes": "Understanding Arabic text and generating human-like responses is a challenging task.", "metrics": null }, { "id": "on-the-robustness-of-arabic-speech-dialect-identification", "name": "On the Robustness of Arabic Speech Dialect Identification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "dialect-id" ], "links": { "paper": "https://arxiv.org/abs/2306.03789" }, "year": 2023, "venue": "Interspeech", "citations": 18, "notes": "As these pipelines require application of ADI tools to potentially out-of-domain data, we aim to investigate how vulnerable the tools", "metrics": null }, { "id": "camelparser2-0-a-state-of-the-art-dependency-parser-for-arab", "name": "CamelParser2.0: A State-of-the-Art Dependency Parser for Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "morphology" ], "links": { "paper": "https://aclanthology.org/2023.arabicnlp-1.15/" }, "year": 2023, "venue": "ARABICNLP", "citations": 17, "notes": "We present CamelParser2.0, an open-source Python-based Arabic dependency parser targeting two popular Arabic dependency formalisms, the Columbia Arabic Treebank", "metrics": null }, { "id": "analyzing-multilingual-competency-of-llms-in-multi-turn-inst", "name": "Analyzing Multilingual Competency of LLMs in Multi-Turn Instruction Following: A Case Study of Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2310.14819" }, "year": 2023, "venue": "ARABICNLP", "citations": 15, "notes": "While significant progress has been made in benchmarking Large Language Models (LLMs) across various tasks, there is a lack of comprehensive evaluation of their", "metrics": null }, { "id": "a-benchmark-and-scoring-algorithm-for-enriching-arabic-synonyms", "name": "A Benchmark and Scoring Algorithm for Enriching Arabic Synonyms", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2302.02232", "website": "https://portal.sina.birzeit.edu/synonyms" }, "year": 2023, "venue": "arXiv 2023", "notes": "We present twofold contributions: an algorithm and a benchmark dataset.", "metrics": null }, { "id": "afrisenti-a-twitter-sentiment-analysis-benchmark-for-african-languages", "name": "AfriSenti: A Twitter Sentiment Analysis Benchmark for African Languages", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2302.08956", "website": "https://afrisenti-semeval.github.io" }, "dialects": [ "magh" ], "year": 2023, "venue": "arXiv 2023", "notes": "We introduce AfriSenti, a sentiment analysis benchmark that contains a total of >110,000 tweets in 14 African languages.", "metrics": null }, { "id": "amurd-annotated-arabic-english-receipt-dataset-for-key-information-ext", "name": "AMuRD: Annotated Arabic-English Receipt Dataset for Key Information Extraction and Classification", "type": "paper", "country": "INTL", "org": "update-for-integrated-business-ai", "license": "unknown", "modality": "text", "tasks": [ "embedding", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2309.09800", "github": "https://github.com/Update-For-Integrated-Business-AI/AMuRD" }, "year": 2023, "venue": "arXiv 2023", "notes": "In this paper, we present AMuRD, a novel multilingual human-annotated dataset specifically designed for information extraction from receipts.", "metrics": null }, { "id": "aner-arabic-and-arabizi-named-entity-recognition-using-transformer-bas", "name": "ANER: Arabic and Arabizi Named Entity Recognition using Transformer-Based Approach", "type": "paper", "country": "INTL", "org": "boda", "license": "unknown", "modality": "text", "tasks": [ "ner" ], "links": { "paper": "https://arxiv.org/abs/2308.14669", "hf": "https://huggingface.co/boda/ANER" }, "year": 2023, "venue": "arXiv 2023", "notes": "We present ANER, a web-based named entity recognizer for the Arabic, and Arabizi languages.", "metrics": { "downloads": 43, "likes": 4, "lastModified": "2023-09-18" } }, { "id": "arabic-fine-grained-entity-recognition", "name": "Arabic Fine-Grained Entity Recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "ner", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2310.17333", "website": "https://sina.birzeit.edu/wojood" }, "year": 2023, "venue": "arXiv 2023", "notes": "Traditional NER systems are typically trained to recognize coarse-grained entities.", "metrics": null }, { "id": "arabic-handwritten-text-line-dataset", "name": "Arabic Handwritten Text Line Dataset", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "ocr" ], "links": { "paper": "https://arxiv.org/abs/2312.07573" }, "year": 2023, "venue": "arXiv 2023", "notes": "In this paper, we present a new dataset specifically designed for historical Arabic script in which we annotate position in word level.", "metrics": null }, { "id": "arabic-mini-climategpt-a-climate-change-and-sustainability-tailored-ar", "name": "Arabic Mini-ClimateGPT : A Climate Change and Sustainability Tailored Arabic LLM", "type": "paper", "country": "AE", "org": "MBZUAI", "license": "unknown", "modality": "text", "tasks": [ "chat", "embedding", "instruction-tuning", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2312.09366", "github": "https://github.com/mbzuai-oryx/ClimateGPT" }, "year": 2023, "venue": "arXiv 2023", "notes": "We propose a light-weight Arabic Mini-ClimateGPT that is built on an open-source LLM and is specifically fine-tuned.", "metrics": null }, { "id": "arabicros-ai-powered-arabic-crossword-puzzle-generation-for-educationa", "name": "ArabIcros: AI-Powered Arabic Crossword Puzzle Generation for Educational Applications", "type": "paper", "country": "INTL", "org": "University of Siena", "license": "unknown", "modality": "text", "tasks": [ "crossword", "education" ], "links": { "paper": "https://aclanthology.org/2023.arabicnlp-1.23/" }, "year": 2023, "venue": "ArabicNLP 2023", "notes": "First Arabic crossword puzzle generator driven by LLMs including GPT-4, GPT-3 variants and BERT, aimed at education.", "metrics": null }, { "id": "araieval-shared-task-persuasion-techniques-and-disinformation-detectio", "name": "ArAIEval Shared Task: Persuasion Techniques and Disinformation Detection in Arabic Text", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2311.03179", "website": "https://araieval.gitlab.io" }, "year": 2023, "venue": "arXiv 2023", "notes": "We present an overview of the ArAIEval shared task, organized as part of the first ArabicNLP 2023 conference co-located with EMNLP 2023.", "metrics": null }, { "id": "arbanking77-intent-detection-neural-model-and-a-new-dataset-in-modern", "name": "ArBanking77: Intent Detection Neural Model and a New Dataset in Modern and Dialectical Arabic", "type": "paper", "country": "PS", "org": "SinaLab, Birzeit University", "license": "unknown", "modality": "text", "tasks": [ "intent-detection", "banking" ], "links": { "paper": "https://arxiv.org/abs/2310.19034", "website": "https://sina.birzeit.edu/arbanking77" }, "dialects": [ "msa", "lev" ], "year": 2023, "venue": "arXiv 2023", "notes": "ArBanking77: 31,404 banking intent queries in MSA and Palestinian dialect, arabized from Banking77, plus a neural intent-detection model.", "metrics": null }, { "id": "arcoq-arabic-closest-opposite-questions-dataset", "name": "ARCOQ: Arabic Closest Opposite Questions Dataset", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2310.14384" }, "year": 2023, "venue": "arXiv 2023", "notes": "This paper presents a dataset for closest opposite questions in Arabic language.", "metrics": null }, { "id": "arpanemo-an-open-source-dataset-for-fine-grained-emotion-recognition-i", "name": "ArPanEmo: An Open-Source Dataset for Fine-Grained Emotion Recognition in Arabic Online Content during COVID-19 Pandemic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://arxiv.org/abs/2305.17580" }, "year": 2023, "venue": "arXiv 2023", "notes": "This paper presents the ArPanEmo dataset, a novel dataset for fine-grained emotion recognition of online posts in Arabic.", "metrics": null }, { "id": "artrivia-harvesting-arabic-wikipedia-to-build-a-new-arabic-question-an", "name": "ArTrivia: Harvesting Arabic Wikipedia to Build A New Arabic Question Answering Dataset", "type": "paper", "country": "INTL", "org": "University of Delaware", "license": "unknown", "modality": "text", "tasks": [ "qa" ], "links": { "paper": "https://aclanthology.org/2023.arabicnlp-1.17/" }, "year": 2023, "venue": "ArabicNLP 2023", "notes": "We present ArTrivia, a new Arabic question-answering dataset consisting of more than 10,000 question-answer pairs along with relevant passages, covering.", "metrics": null }, { "id": "clartts-an-open-source-classical-arabic-text-to-speech-corpus", "name": "ClArTTS: An Open-Source Classical Arabic Text-to-Speech Corpus", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "tts", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2303.00069" }, "dialects": [ "classical" ], "year": 2023, "venue": "arXiv 2023", "notes": "In a move towards filling this gap in resources, we present a speech corpus for Classical Arabic Text-to-Speech.", "metrics": null }, { "id": "cultural-alignment-in-large-language-models-an-explanatory-analysis-ba", "name": "Cultural Alignment in Large Language Models: An Explanatory Analysis Based on Hofstede's Cultural Dimensions", "type": "paper", "country": "INTL", "org": "reemim", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2309.12342", "github": "https://github.com/reemim/Hofstedes_CAT" }, "year": 2023, "venue": "arXiv 2023", "notes": "The deployment of large language models (LLMs) raises concerns regarding their cultural misalignment and potential ramifications on individuals.", "metrics": null }, { "id": "dolphin-a-challenging-and-diverse-benchmark-for-arabic-nlg", "name": "Dolphin: A Challenging and Diverse Benchmark for Arabic NLG", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "qa", "summarization", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2305.14989" }, "year": 2023, "venue": "arXiv 2023", "notes": "We present Dolphin, a novel benchmark that addresses the need for a natural language generation.", "metrics": null }, { "id": "domain-adaptation-for-arabic-machine-translation-the-case-of-financial", "name": "Domain Adaptation for Arabic Machine Translation: The Case of Financial Texts", "type": "paper", "country": "INTL", "org": "Asas AI", "license": "unknown", "modality": "text", "tasks": [ "chat", "translation", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2309.12863", "hf": "https://huggingface.co/asas-ai" }, "year": 2023, "venue": "arXiv 2023", "notes": "Neural machine translation (NMT) has shown impressive performance when trained on large-scale corpora.", "metrics": null }, { "id": "enriching-the-narabizi-treebank-a-multifaceted-approach-to-supporting", "name": "Enriching the NArabizi Treebank: A Multifaceted Approach to Supporting an Under-Resourced Language", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "ner" ], "links": { "paper": "https://arxiv.org/abs/2306.14866" }, "dialects": [ "gulf" ], "year": 2023, "venue": "arXiv 2023", "notes": "We introduce an enriched version of NArabizi Treebank (Seddah et al., 2020) with three main contributions: the addition of two novel annotation layers.", "metrics": null }, { "id": "evaluating-emotion-arcs-across-languages-bridging-the-global-divide-in", "name": "Evaluating Emotion Arcs Across Languages: Bridging the Global Divide in Sentiment Analysis", "type": "paper", "country": "INTL", "org": "dteodore", "license": "unknown", "modality": "text", "tasks": [ "translation", "sentiment", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2306.02213", "github": "https://github.com/dteodore/EmotionArcs" }, "year": 2023, "venue": "arXiv 2023", "notes": "Emotion arcs capture how an individual (or a population) feels over time.", "metrics": null }, { "id": "gari-graph-attention-for-relative-isomorphism-of-arabic-word-embedding", "name": "GARI: Graph Attention for Relative Isomorphism of Arabic Word Embeddings", "type": "paper", "country": "INTL", "org": "asif6827", "license": "unknown", "modality": "text", "tasks": [ "embedding", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2310.13068", "github": "https://github.com/asif6827/GARI" }, "year": 2023, "venue": "arXiv 2023", "notes": "To address this, we propose GARI that combines the distributional training objectives with multiple isomorphism losses guided by the graph attention network.", "metrics": null }, { "id": "having-beer-after-prayer-measuring-cultural-bias-in-large-language-mod", "name": "Having Beer after Prayer? Measuring Cultural Bias in Large Language Models", "type": "paper", "country": "INTL", "org": "tareknaous", "license": "unknown", "modality": "text", "tasks": [ "ner", "sentiment", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2305.14456", "github": "https://github.com/tareknaous/camel" }, "year": 2023, "venue": "arXiv 2023", "notes": "We introduce CAMeL, a novel resource of 628 naturally-occurring prompts and 20,368 entities spanning eight types that contrast Arab and Western cultures.", "metrics": null }, { "id": "hicma-the-handwriting-identification-for-calligraphy-and-manuscripts-i", "name": "HICMA: The Handwriting Identification for Calligraphy and Manuscripts in Arabic Dataset", "type": "paper", "country": "LB", "org": "Lebanese American University", "license": "unknown", "modality": "vision", "tasks": [ "ocr" ], "links": { "paper": "https://aclanthology.org/2023.arabicnlp-1.3/" }, "year": 2023, "venue": "ArabicNLP 2023", "notes": "We present the Handwriting Identification of Manuscripts and Calligraphy in Arabic (HICMA) dataset as the first publicly available dataset with real-world.", "metrics": null }, { "id": "idrisi-d-arabic-and-english-datasets-and-benchmarks-for-location-menti", "name": "IDRISI-D: Arabic and English Datasets and Benchmarks for Location Mention Disambiguation over Disaster Microblogs", "type": "paper", "country": "QA", "org": "Qatar University", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://aclanthology.org/2023.arabicnlp-1.14/" }, "year": 2023, "venue": "ArabicNLP 2023", "notes": "We introduce IDRISI-D, the largest to date English and the first Arabic public LMD datasets.", "metrics": null }, { "id": "impact-of-subword-pooling-strategy-on-cross-lingual-event-detection", "name": "Impact of Subword Pooling Strategy on Cross-lingual Event Detection", "type": "paper", "country": "INTL", "org": "isi-boston", "license": "unknown", "modality": "text", "tasks": [ "ner", "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2302.11365", "github": "https://github.com/isi-boston/ed-pooling" }, "year": 2023, "venue": "arXiv 2023", "notes": "Pre-trained multilingual language models (e.g., mBERT, XLM-RoBERTa) have significantly advanced the state-of-the-art.", "metrics": null }, { "id": "jais-and-jais-chat-arabic-centric-foundation-and", "name": "Jais and Jais-chat: Arabic-Centric Foundation and Instruction-Tuned LLMs", "type": "paper", "country": "INTL", "org": "Various", "license": "apache-2.0", "modality": "text", "year": 2023, "venue": "arXiv 2023", "tasks": [ "pretraining", "instruction-tuning", "llm" ], "links": { "paper": "https://arxiv.org/abs/2308.16149", "hf": "https://huggingface.co/inceptionai/jais-13b-chat" }, "notes": "Arabic-centric bilingual LLMs and chat models trained from scratch.", "metrics": { "downloads": 373, "likes": 168, "lastModified": "2024-09-11" } }, { "id": "larabench-benchmarking-arabic-ai-with-large-language-models", "name": "LAraBench: Benchmarking Arabic AI with Large Language Models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "chat", "asr", "tts", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2305.14982" }, "year": 2023, "venue": "arXiv 2023", "notes": "Recent advancements in Large Language Models (LLMs) have significantly influenced the landscape of language and speech research.", "metrics": null }, { "id": "mavericks-at-nadi-2023-shared-task-unravelling-regional-nuances-throug", "name": "Mavericks at NADI 2023 Shared Task: Unravelling Regional Nuances through Dialect Identification using Transformer-based Approach", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "translation", "dialect-id", "pretraining", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2311.18739" }, "dialects": [ "mixed" ], "year": 2023, "venue": "arXiv 2023", "notes": "In this paper, we present our approach for the \"Nuanced Arabic Dialect Identification (NADI) Shared Task 2023\".", "metrics": null }, { "id": "myvoice-arabic-speech-resource-collaboration-platform", "name": "MyVoice: Arabic Speech Resource Collaboration Platform", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "speech" ], "links": { "paper": "https://arxiv.org/abs/2308.02503" }, "year": 2023, "venue": "arXiv 2023", "notes": "We introduce MyVoice, a crowdsourcing platform designed to collect Arabic speech to enhance dialectal speech technologies.", "metrics": null }, { "id": "n-shot-benchmarking-of-whisper-on-diverse-arabic-speech-recognition", "name": "N-Shot Benchmarking of Whisper on Diverse Arabic Speech Recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "whisper", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2306.02902" }, "dialects": [ "mixed" ], "year": 2023, "venue": "arXiv 2023", "notes": "Evaluates Whisper on diverse Arabic speech recognition benchmarks across dialects and conditions, with n-shot prompting experiments.", "metrics": null }, { "id": "nadi-2023-the-fourth-nuanced-arabic-dialect-identification-shared-task", "name": "NADI 2023: The Fourth Nuanced Arabic Dialect Identification Shared Task", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation", "dialect-id" ], "links": { "paper": "https://arxiv.org/abs/2310.16117" }, "dialects": [ "msa", "mixed" ], "year": 2023, "venue": "arXiv 2023", "notes": "We describe the findings of the fourth Nuanced Arabic Dialect Identification Shared Task (NADI 2023).", "metrics": null }, { "id": "noor-ghateh-a-benchmark-dataset-for-evaluating-arabic-word-segmenters", "name": "Noor-Ghateh: A Benchmark Dataset for Evaluating Arabic Word Segmenters in Hadith Domain", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation", "hadith" ], "links": { "paper": "https://arxiv.org/abs/2307.09630" }, "dialects": [ "classical" ], "year": 2023, "venue": "arXiv 2023", "notes": "We present a standard dataset for analyzing the Arabic segmentation tools.", "metrics": null }, { "id": "osn-mdad-machine-translation-dataset-for-arabic-multi-dialectal-conver", "name": "OSN-MDAD: Machine Translation Dataset for Arabic Multi-Dialectal Conversations on Online Social Media", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "chat", "translation" ], "links": { "paper": "https://arxiv.org/abs/2309.12137" }, "dialects": [ "gulf", "lev", "iraqi", "yemeni", "msa", "mixed" ], "year": 2023, "venue": "arXiv 2023", "notes": "While resources for English language are fairly sufficient to understand content on social media, similar resources in Arabic are still immature.", "metrics": null }, { "id": "pali-a-language-identification-benchmark-for-perso-arabic-scripts", "name": "PALI: A Language Identification Benchmark for Perso-Arabic Scripts", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2304.01322" }, "year": 2023, "venue": "arXiv 2023", "notes": "The Perso-Arabic scripts are a family of scripts that are widely adopted and used by various linguistic communities around the globe.", "metrics": null }, { "id": "readme-benchmarking-multilingual-language-models-for-multi-domain-read", "name": "ReadMe++: Benchmarking Multilingual Language Models for Multi-Domain Readability Assessment", "type": "paper", "country": "INTL", "org": "tareknaous", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2305.14463", "github": "https://github.com/tareknaous/readme" }, "year": 2023, "venue": "arXiv 2023", "notes": "We present a comprehensive evaluation of large language models for multilingual readability assessment.", "metrics": null }, { "id": "rgb-arabic-alphabets-sign-language-dataset", "name": "RGB Arabic Alphabets Sign Language Dataset", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "vision" ], "links": { "paper": "https://arxiv.org/abs/2301.11932" }, "year": 2023, "venue": "arXiv 2023", "notes": "This paper introduces the RGB Arabic Alphabet Sign Language (AASL) dataset.", "metrics": null }, { "id": "salma-arabic-sense-annotated-corpus-and-wsd-benchmarks", "name": "SALMA: Arabic Sense-Annotated Corpus and WSD Benchmarks", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "ner", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2310.19029", "website": "https://sina.birzeit.edu/salma" }, "year": 2023, "venue": "arXiv 2023", "notes": "SALMA, the first Arabic sense-annotated corpus, consists of 34K tokens, which are all sense-annotated.", "metrics": null }, { "id": "semeval-2023-task-12-sentiment-analysis-for-african-languages-afrisent", "name": "SemEval-2023 Task 12: Sentiment Analysis for African Languages (AfriSenti-SemEval)", "type": "paper", "country": "INTL", "org": "afrisenti-semeval", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://arxiv.org/abs/2304.06845", "github": "https://github.com/afrisenti-semeval/afrisent-semeval-2023" }, "dialects": [ "magh" ], "year": 2023, "venue": "arXiv 2023", "notes": "AfriSenti-SemEval: tweet sentiment shared task in 14 African languages, including Algerian and Moroccan Arabic.", "metrics": null }, { "id": "sentiment-analysis-dataset-in-moroccan-dialect-bridging-the-gap-betwee", "name": "Sentiment Analysis Dataset in Moroccan Dialect: Bridging the Gap Between Arabic and Latin Scripted dialect", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://arxiv.org/abs/2303.15987" }, "dialects": [ "magh", "mixed" ], "year": 2023, "venue": "arXiv 2023", "notes": "Sentiment analysis, the automated process of determining emotions or opinions expressed in text.", "metrics": null }, { "id": "taqyim-evaluating-arabic-nlp-tasks-using-chatgpt-models", "name": "Taqyim: Evaluating Arabic NLP Tasks Using ChatGPT Models", "type": "paper", "country": "SA", "org": "ARBML", "license": "unknown", "modality": "text", "tasks": [ "chat", "translation", "diacritization", "sentiment", "summarization", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2306.16322", "github": "https://github.com/ARBML/Taqyim" }, "year": 2023, "venue": "arXiv 2023", "notes": "Additionally, we introduce a new Python interface https://github.com/ARBML/Taqyim that facilitates the evaluation of these tasks effortlessly.", "metrics": null }, { "id": "the-saudi-privacy-policy-dataset", "name": "The Saudi Privacy Policy Dataset", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "privacy", "legal", "classification" ], "links": { "paper": "https://arxiv.org/abs/2304.02757" }, "year": 2023, "venue": "arXiv 2023", "notes": "Saudi Privacy Policy Dataset: 1,000 Saudi websites' Arabic privacy policies annotated against the ten PDPL principles.", "metrics": null }, { "id": "towards-arabic-multimodal-dataset-for-sentiment-analysis", "name": "Towards Arabic Multimodal Dataset for Sentiment Analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "sentiment" ], "links": { "paper": "https://arxiv.org/abs/2306.06322" }, "dialects": [ "msa" ], "year": 2023, "venue": "arXiv 2023", "notes": "Multimodal Sentiment Analysis (MSA) has recently become a centric research direction for many real-world applications.", "metrics": null }, { "id": "toxic-language-detection-a-systematic-review-of-arabic-datasets", "name": "Toxic language detection: a systematic review of Arabic datasets", "type": "paper", "country": "INTL", "org": "imene1", "license": "unknown", "modality": "text", "tasks": [ "survey" ], "links": { "paper": "https://arxiv.org/abs/2312.07228", "github": "https://github.com/Imene1/Arabic-toxic-language" }, "year": 2023, "venue": "arXiv 2023", "notes": "The detection of toxic language in the Arabic language has emerged as an active area of research in recent years.", "metrics": null }, { "id": "using-lstm-and-gru-with-a-new-dataset-for-named-entity-recognition-in", "name": "Using LSTM and GRU With a New Dataset for Named Entity Recognition in the Arabic Language", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "ner" ], "links": { "paper": "https://arxiv.org/abs/2304.03399" }, "year": 2023, "venue": "arXiv 2023", "notes": "In addition, this work proposes long short term memory (LSTM) units and Gated Recurrent Units.", "metrics": null }, { "id": "violet-a-vision-language-model-for-arabic-image-captioning-with-gemini", "name": "Violet: A Vision-Language Model for Arabic Image Captioning with Gemini Decoder", "type": "paper", "country": "INTL", "org": "The University of British C", "license": "unknown", "modality": "multimodal", "tasks": [ "image-captioning" ], "links": { "paper": "https://aclanthology.org/2023.arabicnlp-1.1/" }, "year": 2023, "venue": "ArabicNLP 2023", "notes": "Vision-language model for Arabic image captioning that uses a Gemini decoder.", "metrics": null }, { "id": "wojoodner-2023-the-first-arabic-named-entity-recognition-shared-task", "name": "WojoodNER 2023: The First Arabic Named Entity Recognition Shared Task", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "ner" ], "links": { "paper": "https://arxiv.org/abs/2310.16153" }, "year": 2023, "venue": "arXiv 2023", "notes": "We present WojoodNER-2023, the first Arabic Named Entity Recognition (NER) Shared Task.", "metrics": null }, { "id": "yet-another-model-for-arabic-dialect-identification", "name": "Yet Another Model for Arabic Dialect Identification", "type": "paper", "country": "AE", "org": "MBZUAI", "license": "unknown", "modality": "speech", "tasks": [ "dialect-id" ], "links": { "paper": "https://aclanthology.org/2023.arabicnlp-1.37/" }, "year": 2023, "venue": "ArabicNLP 2023", "notes": "We describe a spoken Arabic dialect identification (ADI) model for Arabic that consistently outperforms previously published results on two benchmark.", "metrics": null }, { "id": "bert-models-for-arabic-text-classification-a-systematic-revi", "name": "BERT Models for Arabic Text Classification: A Systematic Review", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "classification", "pretraining" ], "links": { "paper": "https://doi.org/10.3390/app12115720" }, "year": 2022, "venue": "Applied Sciences", "citations": 143, "notes": "Bidirectional Encoder Representations from Transformers (BERT) has gained increasing attention from researchers and practitioners as it has proven to be an", "metrics": null }, { "id": "semeval-2022-task-6-isarcasmeval-intended-sarcasm-detection", "name": "SemEval-2022 Task 6: iSarcasmEval, Intended Sarcasm Detection in English and Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "classification" ], "links": { "paper": "https://aclanthology.org/2022.semeval-1.111/" }, "year": 2022, "venue": "International Workshop on Semantic Evaluation", "citations": 111, "notes": "iSarcasmEval is the first shared task to target intended sarcasm detection: the data for this task was provided and labelled by the authors of the texts", "metrics": null }, { "id": "arabic-fake-news-detection-based-on-deep-contextualized-embe", "name": "Arabic fake news detection based on deep contextualized embedding models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "classification" ], "links": { "paper": "https://arxiv.org/abs/2205.03114" }, "year": 2022, "venue": "Neural computing & applications (Print)", "citations": 84, "notes": "Social media is becoming a source of news for many people due to its ease and freedom of use.", "metrics": null }, { "id": "arabart-a-pretrained-arabic-sequence-to-sequence-model-for-a", "name": "AraBART: a Pretrained Arabic Sequence-to-Sequence Model for Abstractive Summarization", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "summarization", "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2203.10945" }, "year": 2022, "venue": "Workshop on Arabic Natural Language Processing", "citations": 83, "notes": "Like most natural language understanding and generation tasks, state-of-the-art models for summarization are transformer-based sequence-to-sequence", "metrics": null }, { "id": "sign-language-recognition-for-arabic-alphabets-using-transfe", "name": "Sign Language Recognition for Arabic Alphabets Using Transfer Learning Technique", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "multimodal" ], "links": { "paper": "https://doi.org/10.1155/2022/4567989" }, "year": 2022, "venue": "Computational Intelligence and Neuroscience", "citations": 82, "notes": "Sign language is essential for deaf and mute people to communicate with normal people and themselves.", "metrics": null }, { "id": "arabic-fake-news-detection-based-on-textual-analysis", "name": "Arabic Fake News Detection Based on Textual Analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification" ], "links": { "paper": "https://doi.org/10.1007/s13369-021-06449-y" }, "year": 2022, "venue": "The Arabian journal for science and engineering", "citations": 78, "notes": "Over the years, social media has had a considerable impact on the way we share information and send messages.", "metrics": null }, { "id": "abstractive-arabic-text-summarization-based-on-deep-learning", "name": "Abstractive Arabic Text Summarization Based on Deep Learning", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "summarization" ], "links": { "paper": "https://doi.org/10.1155/2022/1566890" }, "year": 2022, "venue": "Computational Intelligence and Neuroscience", "citations": 77, "notes": "Text summarization (TS) is considered one of the most difficult tasks in natural language processing (NLP).", "metrics": null }, { "id": "a-comprehensive-review-of-arabic-text-summarization", "name": "A Comprehensive Review of Arabic Text Summarization", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "summarization" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2022.3163292" }, "year": 2022, "venue": "IEEE Access", "citations": 72, "notes": "The explosion of online and offline data has changed how we gather, evaluate, and understand data.", "metrics": null }, { "id": "arabic-fake-news-detection-using-deep-learning", "name": "Arabic Fake News Detection Using Deep Learning", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification" ], "links": { "paper": "https://doi.org/10.32604/cmc.2022.021449" }, "year": 2022, "venue": "Computers Materials & Continua", "citations": 71, "notes": ": Nowadays, an unprecedented number of users interact through social media platforms and generate a massive amount of content due to the explosion of online", "metrics": null }, { "id": "heterogeneous-ensemble-deep-learning-model-for-enhanced-arab", "name": "Heterogeneous Ensemble Deep Learning Model for Enhanced Arabic Sentiment Analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.3390/s22103707" }, "year": 2022, "venue": "Italian National Conference on Sensors", "citations": 66, "notes": "Sentiment analysis was nominated as a hot research topic a decade ago for its increasing importance in analyzing the people’s opinions extracted from social", "metrics": null }, { "id": "mawqif-a-multi-label-arabic-dataset-for-target-specific-stan", "name": "Mawqif: A Multi-label Arabic Dataset for Target-specific Stance Detection", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "classification", "dataset" ], "links": { "paper": "https://aclanthology.org/2022.wanlp-1.16/" }, "year": 2022, "venue": "Workshop on Arabic Natural Language Processing", "citations": 53, "notes": "Social media platforms are becoming inherent parts of people’s daily life to express opinions and stances toward topics of varying polarities.", "metrics": null }, { "id": "zaebuc-an-annotated-arabic-english-bilingual-writer-corpus", "name": "ZAEBUC: An Annotated Arabic-English Bilingual Writer Corpus", "type": "paper", "country": "AE", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dataset" ], "links": { "paper": "https://aclanthology.org/2022.lrec-1.9/" }, "year": 2022, "venue": "International Conference on Language Resources and Evaluation", "citations": 53, "notes": "We present ZAEBUC, an annotated Arabic-English bilingual writer corpus comprising short essays by first-year university students at Zayed University in the", "metrics": null }, { "id": "armis-the-arabic-misogyny-and-sexism-corpus-with-annotator-s", "name": "ArMIS - The Arabic Misogyny and Sexism Corpus with Annotator Subjective Disagreements", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "dataset" ], "links": { "paper": "https://aclanthology.org/2022.lrec-1.244/" }, "year": 2022, "venue": "International Conference on Language Resources and Evaluation", "citations": 42, "notes": "The use of misogynistic and sexist language has increased in recent years in social media, and is increasing in the Arabic world in reaction to reforms", "metrics": null }, { "id": "hierarchical-aggregation-of-dialectal-data-for-arabic-dialec", "name": "Hierarchical Aggregation of Dialectal Data for Arabic Dialect Identification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id" ], "links": { "paper": "https://aclanthology.org/2022.lrec-1.489/" }, "year": 2022, "venue": "International Conference on Language Resources and Evaluation", "citations": 26, "notes": "Arabic is a collection of dialectal variants that are historically related but significantly different.", "metrics": null }, { "id": "aradepsu-detecting-depression-and-suicidal-ideation-in-arabi", "name": "AraDepSu: Detecting Depression and Suicidal Ideation in Arabic Tweets Using Transformers", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification" ], "links": { "paper": "https://aclanthology.org/2022.wanlp-1.28/" }, "year": 2022, "venue": "Workshop on Arabic Natural Language Processing", "citations": 23, "notes": "Among mental health diseases, depression is one of the most severe, as it often leads to suicide which is the fourth leading cause of death in the Middle East.", "metrics": null }, { "id": "camel-treebank-an-open-multi-genre-arabic-dependency-treeban", "name": "Camel Treebank: An Open Multi-genre Arabic Dependency Treebank", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "morphology" ], "links": { "paper": "https://aclanthology.org/2022.lrec-1.286/" }, "year": 2022, "venue": "International Conference on Language Resources and Evaluation", "citations": 19, "notes": "We present the Camel Treebank (CAMELTB), a 188K word open-source dependency treebank of Modern Standard and Classical Arabic.", "metrics": null }, { "id": "the-effect-of-arabic-dialect-familiarity-on-data-annotation", "name": "The Effect of Arabic Dialect Familiarity on Data Annotation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://aclanthology.org/2022.wanlp-1.39/" }, "year": 2022, "venue": "Workshop on Arabic Natural Language Processing", "citations": 18, "notes": "Data annotation is the foundation of most natural language processing (NLP) tasks.", "metrics": null }, { "id": "armath-a-dataset-for-solving-arabic-math-word-problems", "name": "ArMATH: a Dataset for Solving Arabic Math Word Problems", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dataset" ], "links": { "paper": "https://aclanthology.org/2022.lrec-1.37/", "github": "https://github.com/reem-codes/ArMATH" }, "year": 2022, "venue": "International Conference on Language Resources and Evaluation", "citations": 16, "notes": "This paper studies solving Arabic Math Word Problems by deep learning.", "metrics": null }, { "id": "emoji-sentiment-roles-for-sentiment-analysis-a-case-study-in", "name": "Emoji Sentiment Roles for Sentiment Analysis: A Case Study in Arabic Texts", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://aclanthology.org/2022.wanlp-1.32/" }, "year": 2022, "venue": "Workshop on Arabic Natural Language Processing", "citations": 15, "notes": "Emoji (digital pictograms) are crucial features for textual sentiment analysis.", "metrics": null }, { "id": "a-benchmark-study-of-contrastive-learning-for-arabic-social-meaning", "name": "A Benchmark Study of Contrastive Learning for Arabic Social Meaning", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id", "sentiment", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2210.12314" }, "year": 2022, "venue": "arXiv 2022", "notes": "In this work, we present a comprehensive benchmark study of state-of-the-art supervised CL methods on a wide array of Arabic social meaning tasks.", "metrics": null }, { "id": "a-deep-cnn-architecture-with-novel-pooling-layer-applied-to-two-sudane", "name": "A Deep CNN Architecture with Novel Pooling Layer Applied to Two Sudanese Arabic Sentiment Datasets", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://arxiv.org/abs/2201.12664" }, "dialects": [ "egy", "gulf", "lev", "magh", "sudanese", "msa", "mixed" ], "year": 2022, "venue": "arXiv 2022", "notes": "Arabic sentiment analysis has become an important research field in recent years.", "metrics": null }, { "id": "a-large-and-diverse-arabic-corpus-for-language-modeling", "name": "A Large and Diverse Arabic Corpus for Language Modeling", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining", "evaluation", "vision" ], "links": { "paper": "https://arxiv.org/abs/2201.09227" }, "dialects": [ "lev" ], "year": 2022, "venue": "arXiv 2022", "notes": "Language models (LMs) have introduced a major paradigm shift in Natural Language Processing.", "metrics": null }, { "id": "a-pilot-study-on-the-collection-and-computational-analysis-of-linguist", "name": "A Pilot Study on the Collection and Computational Analysis of Linguistic Differences Amongst Men and Women in a Kuwaiti Arabic WhatsApp Dataset", "type": "paper", "country": "INTL", "org": "University of Sheffield", "license": "unknown", "modality": "text", "tasks": [ "dialogue" ], "links": { "paper": "https://aclanthology.org/2022.wanlp-1.35/" }, "dialects": [ "gulf" ], "year": 2022, "venue": "WANLP 2022", "notes": "This study focuses on the collection and computational analysis of Kuwaiti Arabic (KA), which is considered a low resource dialect, to test different.", "metrics": null }, { "id": "adversarial-text-to-speech-for-low-resource-languages", "name": "Adversarial Text-to-Speech for low-resource languages", "type": "paper", "country": "INTL", "org": "African Institute for Mathematical Sciences", "license": "unknown", "modality": "speech", "tasks": [ "tts" ], "links": { "paper": "https://aclanthology.org/2022.wanlp-1.8/" }, "year": 2022, "venue": "WANLP 2022", "notes": "We propose a new method for training adversarial text-to-speech (TTS) models for low-resource languages using auxiliary data.", "metrics": null }, { "id": "an-end-to-end-ocr-framework-for-robust-arabic-handwriting-recognition", "name": "An End-to-End OCR Framework for Robust Arabic-Handwriting Recognition using a Novel Transformers-based Model and an Innovative 270 Million-Words Multi-Font Corpus of Classical Arabic with Diacritics", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "ocr", "diacritization", "vision" ], "links": { "paper": "https://arxiv.org/abs/2208.11484" }, "dialects": [ "classical" ], "year": 2022, "venue": "arXiv 2022", "notes": "Notably, we propose an end-to-end text recognition approach using Vision Transformers as an encoder, namely BEIT.", "metrics": null }, { "id": "arabgend-gender-analysis-and-inference-on-arabic-twitter", "name": "ArabGend: Gender Analysis and Inference on Arabic Twitter", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "speech" ], "links": { "paper": "https://arxiv.org/abs/2203.00271", "website": "http://anonymous.com" }, "year": 2022, "venue": "arXiv 2022", "notes": "Gender analysis of Twitter can reveal important socio-cultural differences between male and female users.", "metrics": null }, { "id": "arabsign-a-multi-modality-dataset-and-benchmark-for-continuous-arabic", "name": "ArabSign: A Multi-modality Dataset and Benchmark for Continuous Arabic Sign Language Recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2210.03951" }, "year": 2022, "venue": "arXiv 2022", "notes": "To benchmark this dataset, we propose an encoder-decoder model for Continuous ArSL recognition.", "metrics": null }, { "id": "aranpcc-the-arabic-newspaper-covid-19-corpus", "name": "AraNPCC: The Arabic Newspaper COVID-19 Corpus", "type": "paper", "country": "SA", "org": "The National Center for Data Analytics and Artificial Intelligence", "license": "unknown", "modality": "text", "tasks": [ "corpus", "covid" ], "links": { "paper": "https://aclanthology.org/2022.osact-1.4/" }, "year": 2022, "venue": "OSACT 2022", "notes": "Arabic newspaper corpus on COVID-19 covering 2019-2021 from 12 Arab countries, with over 2 billion words.", "metrics": null }, { "id": "arasas-the-open-source-arabic-semantic-tagger", "name": "AraSAS: The Open Source Arabic Semantic Tagger", "type": "paper", "country": "INTL", "org": "Lancaster University", "license": "unknown", "modality": "text", "tasks": [ "semantic-tagging" ], "links": { "paper": "https://aclanthology.org/2022.osact-1.3/", "github": "https://github.com/UCREL/AraSAS" }, "year": 2022, "venue": "OSACT 2022", "notes": "First open-source Arabic semantic analysis tagging system, built on the UCREL Semantic Analysis System.", "metrics": null }, { "id": "arzen-st-a-three-way-speech-translation-corpus-for-code-switched-egypt", "name": "ArzEn-ST: A Three-way Speech Translation Corpus for Code-Switched Egyptian Arabic - English", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "translation", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2211.12000" }, "dialects": [ "egy" ], "year": 2022, "venue": "arXiv 2022", "notes": "We present our work on collecting ArzEn-ST, a code-switched Egyptian Arabic - English Speech Translation Corpus.", "metrics": null }, { "id": "beyond-arabic-software-for-perso-arabic-script-manipulation", "name": "Beyond Arabic: Software for Perso-Arabic Script Manipulation", "type": "paper", "country": "INTL", "org": "Google Research", "license": "unknown", "modality": "text", "tasks": [ "transliteration" ], "links": { "paper": "https://aclanthology.org/2022.wanlp-1.36/" }, "year": 2022, "venue": "WANLP 2022", "notes": "This paper presents an open-source software library that provides a set of finite-state transducer (FST) components and corresponding utilities.", "metrics": null }, { "id": "cross-lingual-transfer-for-low-resource-arabic-language-understanding", "name": "Cross-lingual transfer for low-resource Arabic language understanding", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "ner" ], "links": { "paper": "https://aclanthology.org/2022.wanlp-1.21/" }, "year": 2022, "venue": "WANLP 2022", "notes": "We adopt a BERT-based architecture and pretrain three models using open-source Wikipedia data and large-scale commercial datasets: monolingual:Arabic.", "metrics": null }, { "id": "cs-um6p-at-semeval-2022-task-6-transformer-based-models-for-intended-s", "name": "CS-UM6P at SemEval-2022 Task 6: Transformer-based Models for Intended Sarcasm Detection in English and Arabic", "type": "paper", "country": "INTL", "org": "abdelkadermh", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2206.08415", "github": "https://github.com/AbdelkaderMH/iSarcasmEval" }, "year": 2022, "venue": "arXiv 2022", "notes": "In this paper, we present our participating system to the intended sarcasm detection task in English and Arabic languages.", "metrics": null }, { "id": "dialect-sentiment-identification-in-nuanced-arabic-tweets-using-an-ens", "name": "Dialect & Sentiment Identification in Nuanced Arabic Tweets Using an Ensemble of Prompt-based, Fine-tuned, and Multitask BERT-Based Models", "type": "paper", "country": "EG", "org": "Cairo University", "license": "unknown", "modality": "text", "tasks": [ "dialect-id", "sentiment" ], "links": { "paper": "https://aclanthology.org/2022.wanlp-1.48/" }, "year": 2022, "venue": "WANLP 2022", "notes": "We present our findings and results in the Nuanced Arabic Dialect Identification Shared Task (NADI 2022) for country-level dialect identification.", "metrics": null }, { "id": "exaasc-a-general-target-based-stance-detection-corpus-in-arabic-langua", "name": "ExaASC: A General Target-Based Stance Detection Corpus in Arabic Language", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2204.13979" }, "year": 2022, "venue": "arXiv 2022", "notes": "This work proposes a new method toward Target-based Stance detection by using the stance of replies toward a most important and arguing target in source tweet.", "metrics": null }, { "id": "gulf-arabic-diacritization-guidelines-initial-dataset-and-results", "name": "Gulf Arabic Diacritization: Guidelines, Initial Dataset, and Results", "type": "paper", "country": "AE", "org": "NYU Abu Dhabi", "license": "unknown", "modality": "text", "tasks": [ "diacritization" ], "links": { "paper": "https://aclanthology.org/2022.wanlp-1.33/" }, "dialects": [ "gulf" ], "year": 2022, "venue": "WANLP 2022", "notes": "We introduce a new Gulf Arabic diacritization dataset composed of 19,850 words based on a subset of the Gumar corpus.", "metrics": null }, { "id": "harnessing-multilingual-resources-to-question-answering-in-arabic", "name": "Harnessing Multilingual Resources to Question Answering in Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa", "quran" ], "links": { "paper": "https://arxiv.org/abs/2205.08024" }, "dialects": [ "classical" ], "year": 2022, "venue": "arXiv 2022", "notes": "The goal of the paper is to predict answers to questions given a passage of Qur'an.", "metrics": null }, { "id": "lans-large-scale-arabic-news-summarization-corpus", "name": "LANS: Large-scale Arabic News Summarization Corpus", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "summarization", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2210.13600" }, "year": 2022, "venue": "arXiv 2022", "notes": "We build, LANS, a large-scale and diverse dataset for Arabic Text Summarization task.", "metrics": null }, { "id": "masader-plus-a-new-interface-for-exploring-500-arabic-nlp-datasets", "name": "Masader Plus: A New Interface for Exploring +500 Arabic NLP Datasets", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "nlp" ], "links": { "paper": "https://arxiv.org/abs/2208.00932", "website": "https://arbml.github.io/masader" }, "year": 2022, "venue": "arXiv 2022", "notes": "In this paper, we introduce Masader Plus, a web interface for users to browse Masader.", "metrics": null }, { "id": "nadi-2022-the-third-nuanced-arabic-dialect-identification-shared-task", "name": "NADI 2022: The Third Nuanced Arabic Dialect Identification Shared Task", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id", "sentiment" ], "links": { "paper": "https://arxiv.org/abs/2210.09582" }, "dialects": [ "mixed" ], "year": 2022, "venue": "arXiv 2022", "notes": "We describe findings of the third Nuanced Arabic Dialect Identification Shared Task (NADI 2022).", "metrics": null }, { "id": "natiq-an-end-to-end-text-to-speech-system-for-arabic", "name": "NatiQ: An End-to-end Text-to-Speech System for Arabic", "type": "paper", "country": "QA", "org": "QCRI", "license": "unknown", "modality": "speech", "tasks": [ "tts" ], "links": { "paper": "https://aclanthology.org/2022.wanlp-1.38/" }, "year": 2022, "venue": "WANLP 2022", "notes": "End-to-end Arabic text-to-speech system using Tacotron-based and faster transformer models with encoder-decoder attention.", "metrics": null }, { "id": "new-results-for-the-text-recognition-of-arabic-maghrib-manuscripts-man", "name": "New Results for the Text Recognition of Arabic Maghrib{ī} Manuscripts -- Managing an Under-resourced Script", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "asr", "ocr" ], "links": { "paper": "https://arxiv.org/abs/2211.16147" }, "year": 2022, "venue": "arXiv 2022", "notes": "HTR models development has become a conventional step for digital humanities projects.", "metrics": null }, { "id": "offensive-language-detection-in-under-resourced-algerian-dialectal-ara", "name": "Offensive Language Detection in Under-resourced Algerian Dialectal Arabic Language", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "nlp" ], "links": { "paper": "https://arxiv.org/abs/2203.10024" }, "dialects": [ "magh", "mixed" ], "year": 2022, "venue": "arXiv 2022", "notes": "This paper addresses the problem of detecting the offensive and abusive content in Facebook comments.", "metrics": null }, { "id": "on-the-arabic-dialects-identification-overcoming-challenges-of-geograp", "name": "On The Arabic Dialects’ Identification: Overcoming Challenges of Geographical Similarities Between Arabic dialects and Imbalanced Datasets", "type": "paper", "country": "INTL", "org": "School of Computer Science", "license": "unknown", "modality": "text", "tasks": [ "dialect-id" ], "links": { "paper": "https://aclanthology.org/2022.wanlp-1.49/" }, "year": 2022, "venue": "WANLP 2022", "notes": "We present a solution to tackle subtask 1 (Country-level dialect identification) of the Nuanced Arabic Dialect Identification (NADI) shared task 2022.", "metrics": null }, { "id": "orca-a-challenging-benchmark-for-arabic-language-understanding", "name": "ORCA: A Challenging Benchmark for Arabic Language Understanding", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2212.10758" }, "year": 2022, "venue": "arXiv 2022", "notes": "In this work, we introduce ORCA, a publicly available benchmark for Arabic language understanding evaluation.", "metrics": null }, { "id": "overview-of-osact5-shared-task-on-arabic-offensive-language-and-hate-s", "name": "Overview of OSACT5 Shared Task on Arabic Offensive Language and Hate Speech Detection", "type": "paper", "country": "QA", "org": "QCRI", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "offensive-language", "shared-task" ], "links": { "paper": "https://aclanthology.org/2022.osact-1.20/" }, "year": 2022, "venue": "OSACT 2022", "notes": "This paper provides an overview of the shard task on detecting offensive language, hate speech, and fine-grained hate speech at the fifth workshop.", "metrics": null }, { "id": "overview-of-the-wanlp-2022-shared-task-on-propaganda-detection-in-arab", "name": "Overview of the WANLP 2022 Shared Task on Propaganda Detection in Arabic", "type": "paper", "country": "QA", "org": "QCRI", "license": "unknown", "modality": "text", "tasks": [ "propaganda", "shared-task" ], "links": { "paper": "https://aclanthology.org/2022.wanlp-1.11/" }, "year": 2022, "venue": "WANLP 2022", "notes": "We propose a shared task on detecting propaganda techniques for Arabic textual content.", "metrics": null }, { "id": "sa-7r-a-saudi-dialect-irony-dataset", "name": "Sa‘7r: A Saudi Dialect Irony Dataset", "type": "paper", "country": "INTL", "org": "College of Computer and Information Sciences", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://aclanthology.org/2022.osact-1.7/" }, "dialects": [ "gulf" ], "year": 2022, "venue": "OSACT 2022", "notes": "We present Sa‘7r ساخرthe Saudi irony dataset, and describe our efforts in constructing it.", "metrics": null }, { "id": "the-arabic-ontology-an-arabic-wordnet-with-ontologically-clean-content", "name": "The Arabic Ontology -- An Arabic Wordnet with Ontologically Clean Content", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2205.09664", "website": "https://ontology.birzeit.edu" }, "year": 2022, "venue": "arXiv 2022", "notes": "We present a formal Arabic wordnet built on the basis of a carefully designed ontology hereby referred to as the Arabic Ontology.", "metrics": null }, { "id": "towards-arabic-sentence-simplification-via-classification-and-generati", "name": "Towards Arabic Sentence Simplification via Classification and Generative Approaches", "type": "paper", "country": "INTL", "org": "nouran-khallaf", "license": "unknown", "modality": "text", "tasks": [ "embedding", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2204.09292" }, "dialects": [ "msa" ], "year": 2022, "venue": "arXiv 2022", "notes": "This paper presents an attempt to build a Modern Standard Arabic (MSA) sentence-level simplification system.", "metrics": null }, { "id": "weakly-and-semi-supervised-learning-for-arabic-text-classification-usi", "name": "Weakly and Semi-Supervised Learning for Arabic Text Classification using Monodialectal Language Models", "type": "paper", "country": "INTL", "org": "Center for Integrative Petroleum Research (CIPR)", "license": "unknown", "modality": "text", "tasks": [ "classification" ], "links": { "paper": "https://aclanthology.org/2022.wanlp-1.24/" }, "year": 2022, "venue": "WANLP 2022", "notes": "To that end, we propose a suite of language-agnostic techniques for large-scale data collection, automatic data annotation, and language model training.", "metrics": null }, { "id": "wojood-nested-arabic-named-entity-corpus-and-recognition", "name": "Wojood: Nested Arabic Named Entity Corpus and Recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "year": 2022, "venue": "arXiv 2022", "tasks": [ "ner", "dataset" ], "links": { "paper": "https://arxiv.org/abs/2205.09651" }, "notes": "Nested named entity corpus in MSA and dialect, with a recognition baseline.", "metrics": null }, { "id": "hyperparameter-tuning-for-machine-learning-algorithms-used-f", "name": "Hyperparameter Tuning for Machine Learning Algorithms Used for Arabic Sentiment Analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.3390/informatics8040079" }, "year": 2021, "venue": "Informatics", "citations": 392, "notes": "In this paper, a comprehensive comparative analysis of various", "metrics": null }, { "id": "bert-for-arabic-topic-modeling-an-experimental-study-on-bert", "name": "BERT for Arabic Topic Modeling: An Experimental Study on BERTopic Technique", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification", "pretraining" ], "links": { "paper": "https://doi.org/10.1016/j.procs.2021.05.096" }, "year": 2021, "venue": "International Conference on Arabic Computational Linguistics", "citations": 176, "notes": "Topic modeling is an unsupervised machine learning technique for finding abstract topics in a large collection of documents.", "metrics": null }, { "id": "pre-training-bert-on-arabic-tweets-practical-considerations", "name": "Pre-Training BERT on Arabic Tweets: Practical Considerations", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2102.10684" }, "year": 2021, "venue": "arXiv.org", "citations": 157, "notes": "We pretrained 5 BERT models that differ in the size of their training sets, mixture of formal and informal Arabic, and linguistic preprocessing.", "metrics": null }, { "id": "arabic-aspect-based-sentiment-analysis-using-bidirectional-g", "name": "Arabic aspect based sentiment analysis using bidirectional GRU based models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://arxiv.org/abs/2101.10539" }, "year": 2021, "venue": "Journal of King Saud University: Computer and Information Sciences", "citations": 118, "notes": "Aspect-based Sentiment analysis (ABSA) accomplishes a fine-grained analysis that defines the aspects of a given document or sentence and the sentiments conveyed", "metrics": null }, { "id": "a-comparative-study-of-effective-approaches-for-arabic-senti", "name": "A comparative study of effective approaches for Arabic sentiment analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "evaluation" ], "links": { "paper": "https://doi.org/10.1016/j.ipm.2020.102438" }, "year": 2021, "venue": "Information Processing & Management", "citations": 99, "notes": "Sentiment analysis (SA) is a natural language processing (NLP) application that aims to analyse and identify sentiment within a piece of text.", "metrics": null }, { "id": "detection-of-hate-speech-in-arabic-tweets-using-deep-learnin", "name": "Detection of hate speech in Arabic tweets using deep learning", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification" ], "links": { "paper": "https://doi.org/10.1007/s00530-020-00742-w" }, "year": 2021, "venue": "Multimedia Systems", "citations": 99, "metrics": null }, { "id": "arabic-fake-news-detection-comparative-study-of-neural-netwo", "name": "Arabic Fake News Detection: Comparative Study of Neural Networks and Transformer-Based Approaches", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification", "evaluation" ], "links": { "paper": "https://doi.org/10.1155/2021/5516945" }, "year": 2021, "venue": "Complex", "citations": 94, "notes": "Fake news detection (FND) involves predicting the likelihood that a particular news article (news report, editorial, expose, etc.) is intentionally deceptive.", "metrics": null }, { "id": "a-survey-of-offensive-language-detection-for-the-arabic-lang", "name": "A Survey of Offensive Language Detection for the Arabic Language", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "hate-speech", "classification" ], "links": { "paper": "https://doi.org/10.1145/3421504" }, "year": 2021, "venue": "ACM Trans. Asian Low Resour. Lang. Inf. Process.", "citations": 91, "notes": "The use of offensive language in user-generated content is a serious problem that needs to be addressed with the latest technology.", "metrics": null }, { "id": "multi-label-arabic-text-classification-in-online-social-netw", "name": "Multi-label Arabic text classification in Online Social Networks", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification" ], "links": { "paper": "https://doi.org/10.1016/J.IS.2021.101785" }, "year": 2021, "venue": "Information Systems", "citations": 82, "metrics": null }, { "id": "arabic-offensive-and-hate-speech-detection-using-a-cross-cor", "name": "Arabic Offensive and Hate Speech Detection Using a Cross-Corpora Multi-Task Learning Model", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification", "dataset" ], "links": { "paper": "https://doi.org/10.3390/informatics8040069" }, "year": 2021, "venue": "Informatics", "citations": 81, "notes": "As social media platforms offer a medium for opinion expression, social phenomena such as hatred, offensive language, racism, and all forms of verbal violence", "metrics": null }, { "id": "benchmarking-transformer-based-language-models-for-arabic-se", "name": "Benchmarking Transformer-based Language Models for Arabic Sentiment and Sarcasm Detection", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "classification", "pretraining" ], "links": { "paper": "https://aclanthology.org/2021.wanlp-1.3/" }, "year": 2021, "venue": "Workshop on Arabic Natural Language Processing", "citations": 80, "notes": "In this paper, we evaluate the performance of 24 of these models on Arabic sentiment and sarcasm detection", "metrics": null }, { "id": "fake-news-detection-in-arabic-tweets-during-the-covid-19-pan", "name": "Fake News Detection in Arabic Tweets during the COVID-19 Pandemic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification" ], "links": { "paper": "https://doi.org/10.14569/IJACSA.2021.0120691" }, "year": 2021, "venue": "arXiv 2021", "citations": 80, "notes": "In March 2020, the World Health Organization declared the COVID-19 outbreak to be a pandemic.", "metrics": null }, { "id": "a-morphologically-annotated-corpus-and-a-morphological-analy", "name": "A Morphologically Annotated Corpus and a Morphological Analyzer for Egyptian Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "morphology", "dataset" ], "links": { "paper": "https://doi.org/10.1016/j.procs.2021.05.084" }, "year": 2021, "venue": "International Conference on Arabic Computational Linguistics", "citations": 77, "metrics": null }, { "id": "enhancing-arabic-aspect-based-sentiment-analysis-using-deep", "name": "Enhancing Arabic aspect-based sentiment analysis using deep learning models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1016/J.CSL.2021.101224" }, "year": 2021, "venue": "Computer Speech and Language", "citations": 77, "metrics": null }, { "id": "preprocessing-arabic-text-on-social-media", "name": "Preprocessing Arabic text on social media", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "preprocessing" ], "links": { "paper": "https://doi.org/10.1016/j.heliyon.2021.e06191" }, "year": 2021, "venue": "Heliyon", "citations": 77, "notes": "Currently, social media plays an important role in daily life and routine.", "metrics": null }, { "id": "dziribert-a-pre-trained-language-model-for-the-algerian-dial", "name": "DziriBERT: a Pre-trained Language Model for the Algerian Dialect", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2109.12346" }, "year": 2021, "venue": "arXiv.org", "citations": 73, "notes": "Pre-trained transformers are now the de facto models in Natural Language Processing given their state-of-the-art results in many tasks and languages.", "metrics": null }, { "id": "arabic-question-answering-system-a-survey", "name": "Arabic question answering system: a survey", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "qa" ], "links": { "paper": "https://doi.org/10.1007/s10462-021-10031-1" }, "year": 2021, "venue": "Artificial Intelligence Review", "citations": 71, "metrics": null }, { "id": "aracovid19-mfh-arabic-covid-19-multi-label-fake-news-hate-sp", "name": "AraCOVID19-MFH: Arabic COVID-19 Multi-label Fake News & Hate Speech Detection Dataset", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification", "dataset" ], "links": { "paper": "https://doi.org/10.1016/j.procs.2021.05.086" }, "year": 2021, "venue": "International Conference on Arabic Computational Linguistics", "citations": 70, "notes": "Along with the COVID-19 pandemic, an \"infodemic\" of false and misleading information has emerged and has complicated the COVID-19 response efforts.", "metrics": null }, { "id": "arabic-aspect-based-sentiment-analysis-a-systematic-literatu", "name": "Arabic Aspect-Based Sentiment Analysis: A Systematic Literature Review", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "sentiment" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2021.3127140" }, "year": 2021, "venue": "IEEE Access", "citations": 69, "notes": "Recently sentiment analysis in Arabic has attracted much attention from researchers.", "metrics": null }, { "id": "arabic-natural-language-processing-for-qur-anic-research-a-s", "name": "Arabic natural language processing for Qur’anic research: a systematic review", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "quran" ], "links": { "paper": "https://doi.org/10.1007/s10462-022-10313-2" }, "year": 2021, "venue": "Artificial Intelligence Review", "citations": 69, "notes": "The Qur’an is a fourteen centuries old divine book in Arabic language that is read and followed by almost two billion Muslims globally as their sacred religious", "metrics": null }, { "id": "unsupervised-neural-networks-for-automatic-arabic-text-summa", "name": "Unsupervised neural networks for automatic Arabic text summarization using document clustering and topic modeling", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "summarization", "classification" ], "links": { "paper": "https://doi.org/10.1016/j.eswa.2021.114652" }, "year": 2021, "venue": "Expert systems with applications", "citations": 69, "metrics": null }, { "id": "deep-learning-for-emotion-analysis-in-arabic-tweets", "name": "Deep learning for emotion analysis in Arabic tweets", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1186/s40537-021-00523-w" }, "year": 2021, "venue": "Journal of Big Data", "citations": 64, "notes": "Currently, expressing feelings through social media requires great consideration as an essential part of our lives; besides sharing ideas and thoughts, we share", "metrics": null }, { "id": "evaluating-various-tokenizers-for-arabic-text-classification", "name": "Evaluating Various Tokenizers for Arabic Text Classification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "preprocessing", "classification", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2106.07540" }, "year": 2021, "venue": "Neural Processing Letters", "citations": 64, "notes": "The first step in any NLP pipeline is to split the text into individual tokens.", "metrics": null }, { "id": "systematic-literature-review-of-dialectal-arabic-identificat", "name": "Systematic Literature Review of Dialectal Arabic: Identification and Detection", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "classification" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2021.3059504" }, "year": 2021, "venue": "IEEE Access", "citations": 64, "notes": "This study comes to chart the field by conducting a systematic literature review that is intended to give insight into the most and least popular research", "metrics": null }, { "id": "towards-arabic-aspect-based-sentiment-analysis-a-transfer-le", "name": "Towards Arabic aspect-based sentiment analysis: a transfer learning-based approach", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1007/s13278-021-00794-4" }, "year": 2021, "venue": "Social Network Analysis and Mining", "citations": 64, "metrics": null }, { "id": "a-deep-learning-framework-for-automatic-detection-of-hate-sp", "name": "A Deep Learning Framework for Automatic Detection of Hate Speech Embedded in Arabic Tweets", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification" ], "links": { "paper": "https://doi.org/10.1007/s13369-021-05383-3" }, "year": 2021, "venue": "The Arabian journal for science and engineering", "citations": 63, "metrics": null }, { "id": "arabic-speech-recognition-by-end-to-end-modular-systems-and", "name": "Arabic Speech Recognition by End-to-End, Modular Systems and Human", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr" ], "links": { "paper": "https://arxiv.org/abs/2101.08454" }, "year": 2021, "venue": "Computer Speech and Language", "citations": 63, "notes": "Recent advances in automatic speech recognition (ASR) have achieved accuracy levels comparable to human transcribers, which led researchers to debate if the", "metrics": null }, { "id": "novel-deep-convolutional-neural-network-based-contextual-rec", "name": "Novel Deep Convolutional Neural Network-Based Contextual Recognition of Arabic Handwritten Scripts", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "ocr" ], "links": { "paper": "https://doi.org/10.3390/e23030340" }, "year": 2021, "venue": "Entropy", "citations": 63, "notes": "Offline Arabic Handwriting Recognition (OAHR) has recently become instrumental in the areas of pattern recognition and image processing due to its application", "metrics": null }, { "id": "recognizing-arabic-handwritten-characters-using-deep-learnin", "name": "Recognizing arabic handwritten characters using deep learning and genetic algorithms", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "ocr" ], "links": { "paper": "https://doi.org/10.1007/s11042-021-11185-4" }, "year": 2021, "venue": "Multimedia tools and applications", "citations": 63, "metrics": null }, { "id": "arabic-speech-recognition-using-end-to-end-deep-learning", "name": "Arabic speech recognition using end-to-end deep learning", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr" ], "links": { "paper": "https://doi.org/10.1049/SIL2.12057" }, "year": 2021, "venue": "IET Signal Processing", "citations": 62, "notes": "In this work, the application of state-of-the-art end-to-end deep learning approaches is inves-tigated to build a", "metrics": null }, { "id": "part-of-speech-tagging-for-arabic-tweets-using-crf-and-bi-ls", "name": "Part-of-speech tagging for Arabic tweets using CRF and Bi-LSTM", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "morphology" ], "links": { "paper": "https://doi.org/10.1016/j.csl.2020.101138" }, "year": 2021, "venue": "Computer Speech and Language", "citations": 61, "notes": "Over the past few years, Twitter has experienced massive growth and the volume of its online content has increased rapidly.", "metrics": null }, { "id": "towards-one-model-to-rule-all-multilingual-strategy-for-dial", "name": "Towards One Model to Rule All: Multilingual Strategy for Dialectal Code-Switching Arabic ASR", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr" ], "links": { "paper": "https://arxiv.org/abs/2105.14779" }, "year": 2021, "venue": "Interspeech", "citations": 59, "notes": "With the advent of globalization, there is an increasing demand for multilingual automatic speech recognition (ASR), handling language and dialectal variation", "metrics": null }, { "id": "sarcasm-and-sentiment-detection-in-arabic-tweets-using-bert", "name": "Sarcasm and Sentiment Detection In Arabic Tweets Using BERT-based Models and Data Augmentation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "classification", "pretraining" ], "links": { "paper": "https://aclanthology.org/2021.wanlp-1.38/" }, "year": 2021, "venue": "Workshop on Arabic Natural Language Processing", "citations": 52, "notes": "In this paper, we describe our efforts on the shared task of sarcasm and sentiment detection in Arabic (Abu Farha et al., 2021)", "metrics": null }, { "id": "deep-multi-task-model-for-sarcasm-detection-and-sentiment-an", "name": "Deep Multi-Task Model for Sarcasm Detection and Sentiment Analysis in Arabic Language", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "classification" ], "links": { "paper": "https://arxiv.org/abs/2106.12488" }, "year": 2021, "venue": "Workshop on Arabic Natural Language Processing", "citations": 50, "notes": "While previous research works tackle SA and sarcasm detection separately, this paper introduces an end-to-end deep Multi-Task Learning (MTL) model, allowing", "metrics": null }, { "id": "alue-arabic-language-understanding-evaluation", "name": "ALUE: Arabic Language Understanding Evaluation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://aclanthology.org/2021.wanlp-1.18/" }, "year": 2021, "venue": "Workshop on Arabic Natural Language Processing", "citations": 48, "notes": "We strongly believe thatmany NLU problems in Arabic are especiallypoised to reap the benefits of such models", "metrics": null }, { "id": "domain-adaptation-for-arabic-cross-domain-and-cross-dialect", "name": "Domain Adaptation for Arabic Cross-Domain and Cross-Dialect Sentiment Analysis from Contextualized Word Embedding", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "embedding" ], "links": { "paper": "https://aclanthology.org/2021.naacl-main.226/" }, "year": 2021, "venue": "North American Chapter of the Association for Computational Linguistics", "citations": 42, "notes": "Finetuning deep pre-trained language models has shown state-of-the-art performances on a wide range of Natural Language Processing (NLP) applications.", "metrics": null }, { "id": "bert-transformer-model-for-detecting-arabic-gpt2-auto-genera", "name": "Bert Transformer model for Detecting Arabic GPT2 Auto-Generated Tweets", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification", "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2101.09345" }, "year": 2021, "venue": "Workshop on Arabic Natural Language Processing", "citations": 34, "notes": "During the last two decades, we have progressively turned to the Internet and social media to find news, entertain conversations and share opinion.", "metrics": null }, { "id": "asad-arabic-social-media-analytics-and-understanding-2021", "name": "ASAD: Arabic Social media Analytics and unDerstanding", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "hate-speech" ], "links": { "paper": "https://aclanthology.org/2021.eacl-demos.14/" }, "year": 2021, "venue": "Conference of the European Chapter of the Association for Computational Linguistics", "citations": 32, "notes": "This system demonstration paper describes ASAD: Arabic Social media Analysis and unDerstanding, a suite of seven individual modules that allows users to", "metrics": null }, { "id": "multi-task-learning-using-a-combination-of-contextualised-an", "name": "Multi-task Learning Using a Combination of Contextualised and Static Word Embeddings for Arabic Sarcasm Detection and Sentiment Analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "embedding", "classification" ], "links": { "paper": "https://aclanthology.org/2021.wanlp-1.39/" }, "year": 2021, "venue": "Workshop on Arabic Natural Language Processing", "citations": 29, "notes": "In this study, we exploited this relationship to enhance both tasks by proposing a multi-task learning approach using a combination of static and", "metrics": null }, { "id": "leveraging-offensive-language-for-sarcasm-and-sentiment-dete", "name": "Leveraging Offensive Language for Sarcasm and Sentiment Detection in Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "hate-speech", "classification" ], "links": { "paper": "https://aclanthology.org/2021.wanlp-1.47/" }, "year": 2021, "venue": "Workshop on Arabic Natural Language Processing", "citations": 28, "notes": "We propose two systems that harness knowledge from multiple tasks to improve the performance of the classifier", "metrics": null }, { "id": "a-contextual-word-embedding-for-arabic-sarcasm-detection-wit", "name": "A Contextual Word Embedding for Arabic Sarcasm Detection with Random Forests", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "embedding", "classification" ], "links": { "paper": "https://aclanthology.org/2021.wanlp-1.43/" }, "year": 2021, "venue": "Workshop on Arabic Natural Language Processing", "citations": 24, "notes": "In this paper, we propose a new approach for improving Arabic sarcasm detection", "metrics": null }, { "id": "combining-context-free-and-contextualized-representations-fo", "name": "Combining Context-Free and Contextualized Representations for Arabic Sarcasm Detection and Sentiment Identification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "classification" ], "links": { "paper": "https://arxiv.org/abs/2103.05683" }, "year": 2021, "venue": "Workshop on Arabic Natural Language Processing", "citations": 24, "notes": "Since their inception, transformer-based language models have led to impressive performance gains across multiple natural language processing tasks.", "metrics": null }, { "id": "empathetic-bert2bert-conversational-model-learning-arabic-la", "name": "Empathetic BERT2BERT Conversational Model: Learning Arabic Language Generation with Little Data", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining", "benchmark", "dataset" ], "links": { "paper": "https://arxiv.org/abs/2103.04353" }, "year": 2021, "venue": "Workshop on Arabic Natural Language Processing", "citations": 23, "notes": "Enabling empathetic behavior in Arabic dialogue agents is an important aspect of building human-like conversational models.", "metrics": null }, { "id": "adapting-marbert-for-improved-arabic-dialect-identification", "name": "Adapting MARBERT for Improved Arabic Dialect Identification: Submission to the NADI 2021 Shared Task", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2103.01065" }, "year": 2021, "venue": "Workshop on Arabic Natural Language Processing", "citations": 22, "notes": "In this paper, we tackle the Nuanced Arabic Dialect Identification (NADI) shared task (Abdul-Mageed et al., 2021) and demonstrate state-of-the-art results on", "metrics": null }, { "id": "bert-based-multi-task-model-for-country-and-province-level-m", "name": "BERT-based Multi-Task Model for Country and Province Level MSA and Dialectal Arabic Identification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2106.12495" }, "year": 2021, "venue": "Workshop on Arabic Natural Language Processing", "citations": 21, "notes": "In this paper, we present our deep learning-based system, submitted to the second NADI shared task for country-level and province-level identification of Modern", "metrics": null }, { "id": "improving-arabic-diacritization-with-regularized-decoding-an", "name": "Improving Arabic Diacritization with Regularized Decoding and Adversarial Training", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "diacritization" ], "links": { "paper": "https://aclanthology.org/2021.acl-short.68/" }, "year": 2021, "venue": "Annual Meeting of the Association for Computational Linguistics", "citations": 20, "notes": "Arabic diacritization is a fundamental task for Arabic language processing.", "metrics": null }, { "id": "the-idc-system-for-sentiment-classification-and-sarcasm-dete", "name": "The IDC System for Sentiment Classification and Sarcasm Detection in Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "classification" ], "links": { "paper": "https://aclanthology.org/2021.wanlp-1.48/" }, "year": 2021, "venue": "Workshop on Arabic Natural Language Processing", "citations": 18, "notes": "In this paper we present designated solutions for sentiment classification and sarcasm detection tasks that were introduced as part of a shared task by Abu", "metrics": null }, { "id": "arsarcasm-shared-task-an-ensemble-bert-model-for-sarcasmdete", "name": "ArSarcasm Shared Task: An Ensemble BERT Model for SarcasmDetection in Arabic Tweets", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "classification", "pretraining" ], "links": { "paper": "https://aclanthology.org/2021.wanlp-1.40/" }, "year": 2021, "venue": "Workshop on Arabic Natural Language Processing", "citations": 17, "notes": "In this work, we present our submission of the sub-task1 of the shared task on sarcasm and sentiment detection in Arabic organized by the 6th Workshop for", "metrics": null }, { "id": "machine-learning-based-approach-for-arabic-dialect-identific", "name": "Machine Learning-Based Approach for Arabic Dialect Identification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id" ], "links": { "paper": "https://aclanthology.org/2021.wanlp-1.34/" }, "year": 2021, "venue": "Workshop on Arabic Natural Language Processing", "citations": 17, "notes": "This paper describes our systems submitted to the Second Nuanced Arabic Dialect Identification Shared Task (NADI 2021)", "metrics": null }, { "id": "wanlp-2021-shared-task-towards-irony-and-sentiment-detection", "name": "WANLP 2021 Shared-Task: Towards Irony and Sentiment Detection in Arabic Tweets using Multi-headed-LSTM-CNN-GRU and MaRBERT", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "classification" ], "links": { "paper": "https://aclanthology.org/2021.wanlp-1.37/" }, "year": 2021, "venue": "Workshop on Arabic Natural Language Processing", "citations": 15, "notes": "This paper presents results and main findings in WANLP 2021 shared tasks one and two", "metrics": null }, { "id": "an-enhanced-corpus-for-arabic-newspapers-comments", "name": "An Enhanced Corpus for Arabic Newspapers Comments", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2102.09965" }, "dialects": [ "magh" ], "year": 2021, "venue": "arXiv 2021", "notes": "In this paper, we propose our enhanced approach to create a dedicated corpus for Algerian Arabic newspapers comments.", "metrics": null }, { "id": "an-open-access-nlp-dataset-for-arabic-dialects-data-collection-labelin", "name": "An open access NLP dataset for Arabic dialects : Data collection, labeling, and model construction", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://arxiv.org/abs/2102.11000" }, "dialects": [ "mixed" ], "year": 2021, "venue": "arXiv 2021", "notes": "In this work, we present an open data set of social data content in several Arabic dialects.", "metrics": null }, { "id": "arabic-offensive-language-on-twitter-analysis-and-experiments", "name": "Arabic Offensive Language on Twitter: Analysis and Experiments", "type": "paper", "country": "QA", "org": "QCRI", "license": "unknown", "modality": "text", "tasks": [ "offensive-language" ], "links": { "paper": "https://aclanthology.org/2021.wanlp-1.13/" }, "year": 2021, "venue": "WANLP 2021", "notes": "We focus on building a large Arabic offensive tweet dataset.", "metrics": null }, { "id": "arabic-speech-emotion-recognition-employing-wav2vec2-0-and-hubert-base", "name": "Arabic Speech Emotion Recognition Employing Wav2vec2.0 and HuBERT Based on BAVED Dataset", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "sentiment", "speech" ], "links": { "paper": "https://arxiv.org/abs/2110.04425" }, "year": 2021, "venue": "arXiv 2021", "notes": "This paper introduces a deep learning constructed emotional recognition model for Arabic speech dialogues.", "metrics": null }, { "id": "aracovid19-mfh-arabic-covid-19-multi-label-fake-news-and-hate-speech-d", "name": "AraCOVID19-MFH: Arabic COVID-19 Multi-label Fake News and Hate Speech Detection Dataset", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2105.03143" }, "dialects": [ "lev" ], "year": 2021, "venue": "arXiv 2021", "notes": "This paper releases \"AraCOVID19-MFH\" a manually annotated multi-label Arabic COVID-19 fake news and hate speech detection dataset.", "metrics": null }, { "id": "aracovid19-ssd-arabic-covid-19-sentiment-and-sarcasm-detection-dataset", "name": "AraCOVID19-SSD: Arabic COVID-19 Sentiment and Sarcasm Detection Dataset", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://arxiv.org/abs/2110.01948" }, "year": 2021, "venue": "arXiv 2021", "notes": "Motivated by the emerging need for annotated datasets that tackle these kinds of problems in the context of COVID-19.", "metrics": null }, { "id": "arastance-a-multi-country-and-multi-domain-dataset-of-arabic-stance-de", "name": "AraStance: A Multi-Country and Multi-Domain Dataset of Arabic Stance Detection for Fact Checking", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2104.13559" }, "dialects": [ "lev" ], "year": 2021, "venue": "arXiv 2021", "notes": "We present our new Arabic Stance Detection dataset (AraStance) of 4,063 claim--article pairs from a diverse set of sources.", "metrics": null }, { "id": "arat5-text-to-text-transformers-for-arabic-language-generation", "name": "AraT5: Text-to-Text Transformers for Arabic Language Generation", "type": "paper", "country": "INTL", "org": "UBC-NLP", "license": "unknown", "modality": "text", "tasks": [ "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2109.12068", "github": "https://github.com/UBC-NLP/araT5" }, "dialects": [ "mixed" ], "year": 2021, "venue": "arXiv 2021", "notes": "For evaluation, we introduce a novel benchmark for ARabic language GENeration (ARGEN), covering seven important tasks.", "metrics": null }, { "id": "arbert-marbert-deep-bidirectional-transformers-for-arabic", "name": "ARBERT & MARBERT: Deep Bidirectional Transformers for Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "year": 2021, "venue": "ACL 2021", "tasks": [ "pretraining", "encoder", "dialect-id" ], "links": { "paper": "https://aclanthology.org/2021.acl-long.551/" }, "notes": "ARBERT and MARBERT: MSA and dialect-heavy encoders, plus the ARLUE benchmark.", "metrics": null }, { "id": "asmdd-arabic-speech-mispronunciation-detection-dataset", "name": "ASMDD: Arabic Speech Mispronunciation Detection Dataset", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "speech" ], "links": { "paper": "https://arxiv.org/abs/2111.01136" }, "dialects": [ "egy" ], "year": 2021, "venue": "arXiv 2021", "notes": "The largest dataset of Arabic speech mispronunciation detections in Egyptian dialogues is introduced.", "metrics": null }, { "id": "calliar-an-online-handwritten-dataset-for-arabic-calligraphy", "name": "Calliar: An Online Handwritten Dataset for Arabic Calligraphy", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "ocr" ], "links": { "paper": "https://arxiv.org/abs/2106.10745" }, "year": 2021, "venue": "arXiv 2021", "notes": "Calligraphy is an essential part of the Arabic heritage and culture.", "metrics": null }, { "id": "exploratory-arabic-offensive-language-dataset-analysis", "name": "Exploratory Arabic Offensive Language Dataset Analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "nlp" ], "links": { "paper": "https://arxiv.org/abs/2101.11434" }, "year": 2021, "venue": "arXiv 2021", "notes": "This paper adding more insights towards resources and datasets used in Arabic offensive language research.", "metrics": null }, { "id": "investigations-on-speech-recognition-systems-for-low-resource-dialecta", "name": "Investigations on Speech Recognition Systems for Low-Resource Dialectal Arabic-English Code-Switching Speech", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "chat", "asr", "speech" ], "links": { "paper": "https://arxiv.org/abs/2108.12881" }, "dialects": [ "egy" ], "year": 2021, "venue": "arXiv 2021", "notes": "In this paper, we present our work on code-switched Egyptian Arabic-English automatic speech recognition (ASR).", "metrics": null }, { "id": "let-mi-an-arabic-levantine-twitter-dataset-for-misogynistic-language", "name": "Let-Mi: An Arabic Levantine Twitter Dataset for Misogynistic Language", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2103.10195" }, "dialects": [ "lev" ], "year": 2021, "venue": "arXiv 2021", "notes": "In this paper, we introduce an Arabic Levantine Twitter dataset for Misogynistic language (LeT-Mi) to be the first benchmark dataset for Arabic misogyny.", "metrics": null }, { "id": "masader-metadata-sourcing-for-arabic-text-and-speech-data-resources", "name": "Masader: Metadata Sourcing for Arabic Text and Speech Data Resources", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2110.06744" }, "year": 2021, "venue": "arXiv 2021", "notes": "The NLP pipeline has evolved dramatically in the last few years.", "metrics": null }, { "id": "nadi-2021-the-second-nuanced-arabic-dialect-identification-shared-task", "name": "NADI 2021: The Second Nuanced Arabic Dialect Identification Shared Task", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id" ], "links": { "paper": "https://arxiv.org/abs/2103.08466" }, "dialects": [ "msa", "mixed" ], "year": 2021, "venue": "arXiv 2021", "notes": "We present the findings and results of the Second Nuanced Arabic Dialect Identification Shared Task (NADI 2021).", "metrics": null }, { "id": "new-arabic-medical-dataset-for-diseases-classification", "name": "New Arabic Medical Dataset for Diseases Classification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2106.15236" }, "year": 2021, "venue": "arXiv 2021", "notes": "We introduce a new Arab medical dataset, which includes two thousand medical documents collected from several Arabic medical websites.", "metrics": null }, { "id": "overview-of-the-arabic-sentiment-analysis-2021-competition-at-kaust", "name": "Overview of the Arabic Sentiment Analysis 2021 Competition at KAUST", "type": "paper", "country": "SA", "org": "KAUST", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "shared-task" ], "links": { "paper": "https://arxiv.org/abs/2109.14456", "website": "https://www.kaggle.com/c/arabic-sentiment-analysis-2021-kaust" }, "year": 2021, "venue": "arXiv 2021", "notes": "Overview of the KAUST Arabic Sentiment Analysis Challenge on tweets, built on 55K ASAD tweets with three sentiment classes.", "metrics": null }, { "id": "overview-of-the-wanlp-2021-shared-task-on-sarcasm-and-sentiment-detect", "name": "Overview of the WANLP 2021 Shared Task on Sarcasm and Sentiment Detection in Arabic", "type": "paper", "country": "INTL", "org": "The University of Edinburgh", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "sarcasm", "shared-task" ], "links": { "paper": "https://aclanthology.org/2021.wanlp-1.36/" }, "year": 2021, "venue": "WANLP 2021", "notes": "This paper provides an overview of the WANLP 2021 shared task on sarcasm and sentiment detection in Arabic.", "metrics": null }, { "id": "qasr-qcri-aljazeera-speech-resource-a-large-scale-annotated-arabic-spe", "name": "QASR: QCRI Aljazeera Speech Resource -- A Large Scale Annotated Arabic Speech Corpus", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "dialect-id", "ner", "evaluation", "speech" ], "links": { "paper": "https://arxiv.org/abs/2106.13000" }, "dialects": [ "mixed" ], "year": 2021, "venue": "arXiv 2021", "notes": "We introduce the largest transcribed Arabic speech corpus, QASR, collected from the broadcast domain.", "metrics": null }, { "id": "self-training-pre-trained-language-models-for-zero-and-few-shot-multi", "name": "Self-Training Pre-Trained Language Models for Zero- and Few-Shot Multi-Dialectal Arabic Sequence Labeling", "type": "paper", "country": "INTL", "org": "mohammadkhalifa", "license": "unknown", "modality": "text", "tasks": [ "ner", "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2101.04758", "github": "https://github.com/mohammadKhalifa/zero-shot-arabic-dialects" }, "dialects": [ "msa", "mixed" ], "year": 2021, "venue": "arXiv 2021", "notes": "We propose to self-train pre-trained language models in zero- and few-shot scenarios to improve performance on data-scarce varieties using only resources.", "metrics": null }, { "id": "sexism-detection-the-first-corpus-in-algerian-dialect-with-a-code-swit", "name": "Sexism detection: The first corpus in Algerian dialect with a code-switching in Arabic/ French and English", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "nlp" ], "links": { "paper": "https://arxiv.org/abs/2104.01443" }, "dialects": [ "magh" ], "year": 2021, "venue": "arXiv 2021", "notes": "In this paper, an approach for hate speech detection against women in Arabic community on social media (e.g.", "metrics": null }, { "id": "the-arabic-parallel-gender-corpus-2-0-extensions-and-analyses", "name": "The Arabic Parallel Gender Corpus 2.0: Extensions and Analyses", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation" ], "links": { "paper": "https://arxiv.org/abs/2110.09216" }, "year": 2021, "venue": "arXiv 2021", "notes": "We introduce a new corpus for gender identification and rewriting in contexts involving one or two target users (I and/or You).", "metrics": null }, { "id": "the-interplay-of-variant-size-and-task-type-in-arabic-pre", "name": "The Interplay of Variant, Size, and Task Type in Arabic Pre-trained Language Models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "year": 2021, "venue": "WANLP 2021", "tasks": [ "pretraining", "encoder" ], "links": { "paper": "https://aclanthology.org/2021.wanlp-1.10/" }, "notes": "Studies how variant, size and task type shape Arabic pretrained encoders (CAMeLBERT).", "metrics": null }, { "id": "ul2c-mapping-user-locations-to-countries-on-arabic-twitter", "name": "UL2C: Mapping User Locations to Countries on Arabic Twitter", "type": "paper", "country": "QA", "org": "QCRI", "license": "unknown", "modality": "text", "tasks": [ "dialect-id" ], "links": { "paper": "https://aclanthology.org/2021.wanlp-1.15/" }, "year": 2021, "venue": "WANLP 2021", "notes": "We present the largest manually labeled dataset for mapping user locations on Arabic Twitter to their corresponding countries.", "metrics": null }, { "id": "zero-resource-multi-dialectal-arabic-natural-language-understanding", "name": "Zero-Resource Multi-Dialectal Arabic Natural Language Understanding", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "ner", "sentiment", "pretraining", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2104.06591" }, "dialects": [ "msa", "mixed" ], "year": 2021, "venue": "arXiv 2021", "notes": "To remedy such performance drop, we propose self-training with unlabeled DA data and apply it in the context of named entity recognition.", "metrics": null }, { "id": "arabic-text-classification-using-deep-learning-models", "name": "Arabic text classification using deep learning models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification" ], "links": { "paper": "https://doi.org/10.1016/j.ipm.2019.102121" }, "year": 2020, "venue": "Information Processing & Management", "citations": 238, "notes": "Text classification or categorization is the process of automatically tagging a textual document with most relevant labels or categories.", "metrics": null }, { "id": "a-review-of-sentiment-analysis-research-in-arabic-language", "name": "A review of sentiment analysis research in Arabic language", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "sentiment" ], "links": { "paper": "https://arxiv.org/abs/2005.12240" }, "year": 2020, "venue": "Future generations computer systems", "citations": 204, "notes": "Sentiment analysis is a task of natural language processing that has recently attracted increasing attention.", "metrics": null }, { "id": "araelectra-pre-training-text-discriminators-for-arabic-langu", "name": "AraELECTRA: Pre-Training Text Discriminators for Arabic Language Understanding", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2012.15516" }, "year": 2020, "venue": "Workshop on Arabic Natural Language Processing", "citations": 176, "notes": "Advances in English language representation enabled a more sample-efficient pre-training task by Efficiently Learning an Encoder that Classifies Token", "metrics": null }, { "id": "deep-learning-cnn-lstm-framework-for-arabic-sentiment-analys", "name": "Deep learning CNN–LSTM framework for Arabic sentiment analysis using textual information shared in social networks", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1007/s13278-020-00668-1" }, "year": 2020, "venue": "Social Network Analysis and Mining", "citations": 162, "metrics": null }, { "id": "aragpt2-pre-trained-transformer-for-arabic-language-generati", "name": "AraGPT2: Pre-Trained Transformer for Arabic Language Generation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2012.15520" }, "year": 2020, "venue": "Workshop on Arabic Natural Language Processing", "citations": 161, "notes": "Recently, pre-trained transformer-based architectures have proven to be very efficient at language modeling and understanding, given that they are trained on a", "metrics": null }, { "id": "a-panoramic-survey-of-natural-language-processing-in-the-ara", "name": "A panoramic survey of natural language processing in the Arab world", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey" ], "links": { "paper": "https://arxiv.org/abs/2011.12631" }, "year": 2020, "venue": "Communications of the ACM", "citations": 155, "notes": "The term natural language refers to any system of symbolic communication (spoken, signed or written) without intentional human planning and design.", "metrics": null }, { "id": "nadi-2020-the-first-nuanced-arabic-dialect-identification-sh", "name": "NADI 2020: The First Nuanced Arabic Dialect Identification Shared Task", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2010.11334" }, "year": 2020, "venue": "Workshop on Arabic Natural Language Processing", "citations": 146, "notes": "We present the results and findings of the First Nuanced Arabic Dialect Identification Shared Task (NADI).", "metrics": null }, { "id": "deeparslr-a-novel-signer-independent-deep-learning-framework", "name": "DeepArSLR: A Novel Signer-Independent Deep Learning Framework for Isolated Arabic Sign Language Gestures Recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "multimodal" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2020.2990699" }, "year": 2020, "venue": "IEEE Access", "citations": 127, "notes": "Hand gesture recognition has attracted the attention of many researchers due to its wide applications in robotics, games, virtual reality, sign language and", "metrics": null }, { "id": "a-deep-learning-approach-for-automatic-hate-speech-detection", "name": "A Deep Learning Approach for Automatic Hate Speech Detection in the Saudi Twittersphere", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification" ], "links": { "paper": "https://doi.org/10.3390/app10238614" }, "year": 2020, "venue": "Applied Sciences", "citations": 119, "notes": "With the rise of hate speech phenomena in the Twittersphere, significant research efforts have been undertaken in order to provide automatic solutions for", "metrics": null }, { "id": "arabic-sign-language-recognition-and-generating-arabic-speec", "name": "Arabic Sign Language Recognition and Generating Arabic Speech Using Convolutional Neural Network", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "multimodal" ], "links": { "paper": "https://doi.org/10.1155/2020/3685614" }, "year": 2020, "venue": "Wireless Communications and Mobile Computing", "citations": 117, "notes": "Sign language encompasses the movement of the arms and hands as a means of communication for people with hearing disabilities.", "metrics": null }, { "id": "qadi-arabic-dialect-identification-in-the-wild", "name": "QADI: Arabic Dialect Identification in the Wild", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id" ], "links": { "paper": "https://arxiv.org/abs/2005.06557" }, "year": 2020, "venue": "Workshop on Arabic Natural Language Processing", "citations": 108, "notes": "In this paper, we present a method for rapidly constructing a tweet dataset containing a wide range of country-level Arabic dialects —covering 18 different", "metrics": null }, { "id": "intelligent-detection-of-hate-speech-in-arabic-social-networ", "name": "Intelligent detection of hate speech in Arabic social network: A machine learning approach", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification" ], "links": { "paper": "https://doi.org/10.1177/0165551520917651" }, "year": 2020, "venue": "Journal of information science", "citations": 106, "notes": "Nowadays, cyber hate speech is increasingly growing, which forms a serious problem worldwide by threatening the cohesion of civil societies.", "metrics": null }, { "id": "hate-and-offensive-speech-detection-on-arabic-social-media", "name": "Hate and offensive speech detection on Arabic social media", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification" ], "links": { "paper": "https://doi.org/10.1016/j.osnem.2020.100096" }, "year": 2020, "venue": "Online Soc. Networks Media", "citations": 102, "notes": "We are witnessing an increasing proliferation of hate speech on social media targeting individuals for their protected characteristics.", "metrics": null }, { "id": "deep-learning-for-arabic-subjective-sentiment-analysis-chall", "name": "Deep learning for Arabic subjective sentiment analysis: Challenges and research opportunities", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1016/j.asoc.2020.106836" }, "year": 2020, "venue": "Applied Soft Computing", "citations": 101, "notes": "The fields of machine learning and Web technologies have witnessed significant development in the last years.", "metrics": null }, { "id": "extractive-arabic-text-summarization-using-modified-pagerank", "name": "Extractive Arabic Text Summarization Using Modified PageRank Algorithm", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "summarization" ], "links": { "paper": "https://doi.org/10.1016/j.eij.2019.11.001" }, "year": 2020, "venue": "Egyptian Informatics Journal", "citations": 98, "notes": "This paper proposed an approach for Arabic text summarization.", "metrics": null }, { "id": "deep-bidirectional-lstm-network-learning-based-sentiment-ana", "name": "Deep Bidirectional LSTM Network Learning-Based Sentiment Analysis for Arabic Text", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1515/jisys-2020-0021" }, "year": 2020, "venue": "Journal of Intelligent Systems", "citations": 96, "notes": "Sentiment analysis aims to predict sentiment polarities (positive, negative or neutral) of a given piece of text.", "metrics": null }, { "id": "arabic-sign-language-recognition-through-deep-neural-network", "name": "Arabic Sign Language Recognition through Deep Neural Networks Fine-Tuning", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "multimodal" ], "links": { "paper": "https://doi.org/10.3991/ijoe.v16i05.13087" }, "year": 2020, "venue": "Int. J. Online Biomed. Eng.", "citations": 92, "notes": "Sign Language is considered the main communication tool for deaf or hearing impaired people.", "metrics": null }, { "id": "egyptian-arabic-speech-emotion-recognition-using-prosodic-sp", "name": "Egyptian Arabic speech emotion recognition using prosodic, spectral and wavelet features", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1016/j.specom.2020.04.005" }, "year": 2020, "venue": "Speech Communication", "citations": 90, "notes": "Speech emotion recognition (SER) has recently been receiving increased interest due to the rapid advancements in affective computing and human computer", "metrics": null }, { "id": "a-sentiment-analysis-approach-to-predict-an-individual-s-awa", "name": "A Sentiment Analysis Approach to Predict an Individual’s Awareness of the Precautionary Procedures to Prevent COVID-19 Outbreaks in Saudi Arabia", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.3390/ijerph18010218" }, "year": 2020, "venue": "International Journal of Environmental Research and Public Health", "citations": 85, "notes": "In March 2020, the World Health Organization (WHO) declared the outbreak of Coronavirus disease 2019 (COVID-19) as a pandemic, which affected all countries", "metrics": null }, { "id": "impact-of-stemming-and-word-embedding-on-deep-learning-based", "name": "Impact of Stemming and Word Embedding on Deep Learning-Based Arabic Text Categorization", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "preprocessing", "classification" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2020.3009217" }, "year": 2020, "venue": "IEEE Access", "citations": 81, "notes": "Document classification is a classical problem in information retrieval, and plays an important role in a variety of applications.", "metrics": null }, { "id": "comparative-performance-of-machine-learning-and-deep-learnin", "name": "Comparative Performance of Machine Learning and Deep Learning Algorithms for Arabic Hate Speech Detection in OSNs", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification", "evaluation" ], "links": { "paper": "https://doi.org/10.1007/978-3-030-44289-7_24" }, "year": 2020, "venue": "International Conferences on Artificial Intelligence and Computer Vision", "citations": 76, "metrics": null }, { "id": "arabic-sentiment-analysis-a-systematic-literature-review", "name": "Arabic Sentiment Analysis: A Systematic Literature Review", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "sentiment" ], "links": { "paper": "https://doi.org/10.1155/2020/7403128" }, "year": 2020, "venue": "Applied Computational Intelligence and Soft Computing", "citations": 75, "notes": "This paper introduces a systematic review of the existing literature relevant to ASA.", "metrics": null }, { "id": "arabic-text-summarization-using-deep-learning-approach", "name": "Arabic text summarization using deep learning approach", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "summarization" ], "links": { "paper": "https://doi.org/10.1186/s40537-020-00386-7" }, "year": 2020, "venue": "Journal of Big Data", "citations": 70, "notes": "Natural language processing has witnessed remarkable progress with the advent of deep learning techniques.", "metrics": null }, { "id": "a-systematic-review-of-text-classification-research-based-on", "name": "A systematic review of text classification research based on deep learning models in Arabic language", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "classification" ], "links": { "paper": "https://doi.org/10.11591/IJECE.V10I6.PP6629-6643" }, "year": 2020, "venue": "arXiv 2020", "citations": 69, "notes": "This paper undertakes a systematic review of the latest research in the field of the classification of Arabic texts.", "metrics": null }, { "id": "hate-speech-detection-using-word-embedding-and-deep-learning", "name": "Hate Speech Detection using Word Embedding and Deep Learning in the Arabic Language Context", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "embedding", "classification" ], "links": { "paper": "https://doi.org/10.5220/0008954004530460" }, "year": 2020, "venue": "International Conference on Pattern Recognition Applications and Methods", "citations": 68, "notes": ": Hate speech over online social networks is a worldwide problem that leads for diminishing the cohesion of civil societies.", "metrics": null }, { "id": "detection-of-hate-speech-in-covid-19-related-tweets-in-the-a", "name": "Detection of Hate Speech in COVID-19–Related Tweets in the Arab Region: Deep Learning and Topic Modeling Approach", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification" ], "links": { "paper": "https://doi.org/10.2196/22609" }, "year": 2020, "venue": "Journal of Medical Internet Research", "citations": 64, "notes": "The massive scale of social media platforms requires an automatic solution for detecting hate speech.", "metrics": null }, { "id": "predicting-depression-symptoms-in-an-arabic-psychological-fo", "name": "Predicting Depression Symptoms in an Arabic Psychological Forum", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "benchmark" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2020.2981834" }, "year": 2020, "venue": "IEEE Access", "citations": 64, "notes": "Recently, social media platforms have been widely used as a communication tool on social networks.", "metrics": null }, { "id": "improving-sentiment-analysis-of-arabic-tweets-by-one-way-ano", "name": "Improving Sentiment Analysis of Arabic Tweets by One-way ANOVA", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1016/j.jksuci.2020.10.023" }, "year": 2020, "venue": "Journal of King Saud University: Computer and Information Sciences", "citations": 62, "notes": "Social media is an indispensable necessity for modern life.", "metrics": null }, { "id": "recent-advances-in-nlp-the-case-of-arabic-language", "name": "Recent Advances in NLP: The Case of Arabic Language", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://doi.org/10.1007/978-3-030-34614-0" }, "year": 2020, "venue": "Studies in Computational Intelligence", "citations": 62, "metrics": null }, { "id": "nabiha-an-arabic-dialect-chatbot", "name": "Nabiha: An Arabic Dialect Chatbot", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://doi.org/10.14569/ijacsa.2020.0110357" }, "year": 2020, "venue": "International Journal of Advanced Computer Science and Applications", "citations": 60, "notes": "Nowadays, we are living in the era of technology and innovation that impact various fields, including sciences.", "metrics": null }, { "id": "habibi-a-multi-dialect-multi-national-arabic-song-lyrics-cor", "name": "Habibi - a multi Dialect multi National Arabic Song Lyrics Corpus", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dataset" ], "links": { "paper": "https://aclanthology.org/2020.lrec-1.165/" }, "year": 2020, "venue": "International Conference on Language Resources and Evaluation", "citations": 52, "notes": "This paper introduces Habibi the first Arabic Song Lyrics corpus", "metrics": null }, { "id": "investigating-the-effects-of-gender-dialect-and-training-siz", "name": "Investigating the effects of gender, dialect, and training size on the performance of Arabic speech recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "evaluation" ], "links": { "paper": "https://doi.org/10.1007/s10579-020-09505-5" }, "year": 2020, "venue": "Language Resources and Evaluation", "citations": 50, "notes": "Research in Arabic automatic speech recognition (ASR) is constrained by datasets of limited size, and of highly variable content and quality.", "metrics": null }, { "id": "daict-a-dialectal-arabic-irony-corpus-extracted-from-twitter", "name": "DAICT: A Dialectal Arabic Irony Corpus Extracted from Twitter", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "dataset" ], "links": { "paper": "https://aclanthology.org/2020.lrec-1.768/" }, "year": 2020, "venue": "International Conference on Language Resources and Evaluation", "citations": 47, "notes": "We query Twitter using irony-related hashtags to collect ironic messages, which are then manually annotated by two linguists according to our working", "metrics": null }, { "id": "empathy-driven-arabic-conversational-chatbot", "name": "Empathy-driven Arabic Conversational Chatbot", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://aclanthology.org/2020.wanlp-1.6/" }, "year": 2020, "venue": "Workshop on Arabic Natural Language Processing", "citations": 38, "notes": "To address these challenges, we create an Arabic conversational dataset that comprises empathetic responses", "metrics": null }, { "id": "a-unified-model-for-arabizi-detection-and-transliteration-us", "name": "A Unified Model for Arabizi Detection and Transliteration using Sequence-to-Sequence Models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "preprocessing", "classification" ], "links": { "paper": "https://aclanthology.org/2020.wanlp-1.15/" }, "year": 2020, "venue": "Workshop on Arabic Natural Language Processing", "citations": 35, "notes": "In this paper, we present the first effort on a unified model for Arabizi detection and transliteration into a code-mixed output with consistent Arabic", "metrics": null }, { "id": "language-resources-for-maghrebi-arabic-dialects-nlp-a-survey", "name": "Language resources for Maghrebi Arabic dialects’ NLP: a survey", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "dataset" ], "links": { "paper": "https://doi.org/10.1007/s10579-020-09490-9" }, "year": 2020, "venue": "Language Resources and Evaluation", "citations": 34, "metrics": null }, { "id": "hate-speech-detection-in-saudi-twittersphere-a-deep-learning", "name": "Hate Speech Detection in Saudi Twittersphere: A Deep Learning Approach", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification" ], "links": { "paper": "https://aclanthology.org/2020.wanlp-1.2/" }, "year": 2020, "venue": "Workshop on Arabic Natural Language Processing", "citations": 29, "notes": "This paper, therefore, aims to investigate several neural network models based on Convolutional Neural Network (CNN) and Recurrent Neural Networks (RNN) to", "metrics": null }, { "id": "is-it-great-or-terrible-preserving-sentiment-in-neural-machi", "name": "Is it Great or Terrible? Preserving Sentiment in Neural Machine Translation of Arabic Reviews", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "translation" ], "links": { "paper": "https://arxiv.org/abs/2010.13814" }, "year": 2020, "venue": "Workshop on Arabic Natural Language Processing", "citations": 24, "notes": "Since the advent of Neural Machine Translation (NMT) approaches there has been a tremendous improvement in the quality of automatic translation.", "metrics": null }, { "id": "morphological-analysis-and-disambiguation-for-gulf-arabic-th", "name": "Morphological Analysis and Disambiguation for Gulf Arabic: The Interplay between Resources and Methods", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "morphology", "dataset" ], "links": { "paper": "https://aclanthology.org/2020.lrec-1.480/" }, "year": 2020, "venue": "International Conference on Language Resources and Evaluation", "citations": 24, "notes": "In this paper we present the first full morphological analysis and disambiguation system for Gulf Arabic", "metrics": null }, { "id": "weighted-combination-of-bert-and-n-gram-features-for-nuanced", "name": "Weighted combination of BERT and N-GRAM features for Nuanced Arabic Dialect Identification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id", "pretraining" ], "links": { "paper": "https://aclanthology.org/2020.wanlp-1.27/" }, "year": 2020, "venue": "Workshop on Arabic Natural Language Processing", "citations": 24, "notes": "In this paper, we investigate the Arabic dialect identification task, from two perspectives: country-level dialect identification from 21 Arab countries, and", "metrics": null }, { "id": "a-spelling-correction-corpus-for-multiple-arabic-dialects", "name": "A Spelling Correction Corpus for Multiple Arabic Dialects", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "preprocessing", "dataset" ], "links": { "paper": "https://aclanthology.org/2020.lrec-1.508/" }, "year": 2020, "venue": "International Conference on Language Resources and Evaluation", "citations": 23, "notes": "In this paper, we present the MADAR CODA Corpus, a collection of 10,000 sentences from five Arabic city dialects (Beirut, Cairo, Doha, Rabat, and Tunis)", "metrics": null }, { "id": "on-the-importance-of-tokenization-in-arabic-embedding-models", "name": "On the Importance of Tokenization in Arabic Embedding Models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "preprocessing" ], "links": { "paper": "https://aclanthology.org/2020.wanlp-1.11/" }, "year": 2020, "venue": "Workshop on Arabic Natural Language Processing", "citations": 23, "notes": "In this work, we propose two embedding strategies that modify the tokenization phase of traditional word embedding models (Word2Vec) and contextual word", "metrics": null }, { "id": "effects-of-dialectal-code-switching-on-speech-modules-a-stud", "name": "Effects of Dialectal Code-Switching on Speech Modules: A Study Using Egyptian Arabic Broadcast Speech", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "asr", "dataset" ], "links": { "paper": "https://doi.org/10.21437/interspeech.2020-2271" }, "year": 2020, "venue": "Interspeech", "citations": 22, "notes": "The intra-utterance code-switching (CS) is defined as the alternation between two or more languages within the same ut-terance.", "metrics": null }, { "id": "a-semi-supervised-bert-approach-for-arabic-named-entity-reco", "name": "A Semi-Supervised BERT Approach for Arabic Named Entity Recognition", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "ner", "pretraining" ], "links": { "paper": "https://aclanthology.org/2020.wanlp-1.5/" }, "year": 2020, "venue": "Workshop on Arabic Natural Language Processing", "citations": 21, "notes": "This paper proposes a semi-supervised learning approach to train a BERT-based NER model using labeled and semi-labeled datasets", "metrics": null }, { "id": "deep-diacritization-efficient-hierarchical-recurrence-for-im", "name": "Deep Diacritization: Efficient Hierarchical Recurrence for Improved Arabic Diacritization", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "diacritization" ], "links": { "paper": "https://arxiv.org/abs/2011.00538" }, "year": 2020, "venue": "Workshop on Arabic Natural Language Processing", "citations": 20, "notes": "We propose a novel architecture for labelling character sequences that achieves state-of-the-art results on the Tashkeela Arabic diacritization benchmark.", "metrics": null }, { "id": "identifying-sentiments-in-algerian-code-switched-user-genera", "name": "Identifying Sentiments in Algerian Code-switched User-generated Comments", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://aclanthology.org/2020.lrec-1.328/" }, "year": 2020, "venue": "International Conference on Language Resources and Evaluation", "citations": 18, "notes": "We present in this paper our work on Algerian language, an under-resourced North African colloquial Arabic variety, for which we built a comparably large", "metrics": null }, { "id": "arabic-dialects-identification-for-all-arabic-countries", "name": "Arabic Dialects Identification for All Arabic countries", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id" ], "links": { "paper": "https://aclanthology.org/2020.wanlp-1.32/" }, "year": 2020, "venue": "Workshop on Arabic Natural Language Processing", "citations": 16, "notes": "In this paper, several techniques with multiple algorithms are applied for Arabic dialects identification starting from removing noise till classification task", "metrics": null }, { "id": "a-benchmark-arabic-dataset-for-commonsense-explanation", "name": "A Benchmark Arabic Dataset for Commonsense Explanation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2012.10251" }, "year": 2020, "venue": "arXiv 2020", "notes": "In this paper, we present a benchmark Arabic dataset for commonsense explanation.", "metrics": null }, { "id": "an-arabic-tweets-sentiment-analysis-dataset-atsad-using-distant-superv", "name": "An Arabic Tweets Sentiment Analysis Dataset (ATSAD) using Distant Supervision and Self Training", "type": "paper", "country": "INTL", "org": "University o", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://aclanthology.org/2020.osact-1.1/" }, "year": 2020, "venue": "OSACT 2020", "notes": "We present an Arabic Sentiment Analysis Corpus collected from Twitter, which contains 36K tweets labelled into positive and negative.", "metrics": null }, { "id": "an-empirical-study-of-pre-trained-transformers-for-arabic-information", "name": "An Empirical Study of Pre-trained Transformers for Arabic Information Extraction", "type": "paper", "country": "INTL", "org": "lanwuwei", "license": "unknown", "modality": "text", "tasks": [ "ner", "pretraining" ], "links": { "paper": "https://arxiv.org/abs/2004.14519", "github": "https://github.com/lanwuwei/GigaBERT" }, "year": 2020, "venue": "arXiv 2020", "notes": "Multilingual pre-trained Transformers, such as mBERT (Devlin et al., 2019) and XLM-RoBERTa.", "metrics": null }, { "id": "araacom-arabic-algerian-corpus-for-opinion-mining", "name": "ARAACOM: ARAbic Algerian Corpus for Opinion Mining", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "vision" ], "links": { "paper": "https://arxiv.org/abs/2001.08010" }, "dialects": [ "magh" ], "year": 2020, "venue": "arXiv 2020", "notes": "In this paper, we propose our approach, for opinion mining in Arabic Algerian news paper.", "metrics": null }, { "id": "arabert-transformer-based-model-for-arabic-language", "name": "AraBERT: Transformer-based Model for Arabic Language Understanding", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "year": 2020, "venue": "OSACT 2020", "tasks": [ "pretraining", "encoder" ], "links": { "paper": "https://aclanthology.org/2020.osact-1.2/", "github": "https://github.com/aub-mind/arabert" }, "notes": "BERT-style Arabic encoder pretrained on a large MSA corpus.", "metrics": null }, { "id": "arabic-dialect-identification-an-arabic-bert-model-with-data-augmentat", "name": "Arabic dialect identification: An Arabic-BERT model with data augmentation and ensembling strategy", "type": "paper", "country": "MA", "org": "SI2M Lab (INSEA)", "license": "unknown", "modality": "text", "tasks": [ "dialect-id" ], "links": { "paper": "https://aclanthology.org/2020.wanlp-1.28/" }, "year": 2020, "venue": "WANLP 2020", "notes": "ArabicProcessors team Arabic-BERT system with data augmentation and ensembling for NADI 2020 country- and province-level dialect identification.", "metrics": null }, { "id": "araweat-multidimensional-analysis-of-biases-in-arabic-word-embeddings", "name": "AraWEAT: Multidimensional Analysis of Biases in Arabic Word Embeddings", "type": "paper", "country": "INTL", "org": "Data and Web Science Research Group", "license": "unknown", "modality": "text", "tasks": [ "embedding" ], "links": { "paper": "https://aclanthology.org/2020.wanlp-1.17/" }, "dialects": [ "egy", "msa" ], "year": 2020, "venue": "WANLP 2020", "notes": "We conduct an extensive analysis of biases in Arabic word embeddings by applying a range of recently introduced bias tests on a variety of embedding spaces.", "metrics": null }, { "id": "arcov-19-the-first-arabic-covid-19-twitter-dataset-with-propagation-ne", "name": "ArCOV-19: The First Arabic COVID-19 Twitter Dataset with Propagation Networks", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "chat", "embedding" ], "links": { "paper": "https://arxiv.org/abs/2004.05861" }, "year": 2020, "venue": "arXiv 2020", "notes": "We present ArCOV-19, an Arabic COVID-19 Twitter dataset that spans one year, covering the period from 27th of January 2020 till 31st of January 2021.", "metrics": null }, { "id": "arcov19-rumors-arabic-covid-19-twitter-dataset-for-misinformation-dete", "name": "ArCOV19-Rumors: Arabic COVID-19 Twitter Dataset for Misinformation Detection", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "chat", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2010.08768" }, "dialects": [ "lev" ], "year": 2020, "venue": "arXiv 2020", "notes": "In this paper we introduce ArCOV19-Rumors, an Arabic COVID-19 Twitter dataset for misinformation detection composed of tweets containing claims.", "metrics": null }, { "id": "asad-a-twitter-based-benchmark-arabic-sentiment-analysis-dataset", "name": "ASAD: A Twitter-based Benchmark Arabic Sentiment Analysis Dataset", "type": "paper", "country": "SA", "org": "KAUST", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "benchmark" ], "links": { "paper": "https://arxiv.org/abs/2011.00578" }, "year": 2020, "venue": "arXiv 2020", "notes": "ASAD: large annotated Twitter benchmark for Arabic sentiment analysis, launched as the KAUST-sponsored competition dataset.", "metrics": null }, { "id": "automatic-arabic-dialect-identification-systems-for-written-texts-a-su", "name": "Automatic Arabic Dialect Identification Systems for Written Texts: A Survey", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "tts", "translation", "dialect-id", "evaluation", "survey", "speech" ], "links": { "paper": "https://arxiv.org/abs/2009.12622" }, "dialects": [ "mixed" ], "year": 2020, "venue": "arXiv 2020", "notes": "In this paper, we present a comprehensive survey of Arabic dialect identification research in written texts.", "metrics": null }, { "id": "contextual-embeddings-for-arabic-english-code-switched-data", "name": "Contextual Embeddings for Arabic-English Code-Switched Data", "type": "paper", "country": "EG", "org": "German University in Cairo", "license": "unknown", "modality": "text", "tasks": [ "code-switching", "embedding" ], "links": { "paper": "https://aclanthology.org/2020.wanlp-1.20/" }, "year": 2020, "venue": "WANLP 2020", "notes": "We propose an open source trained bilingual contextual word embedding models of FLAIR, BERT, and ELECTRA.", "metrics": null }, { "id": "dialex-a-benchmark-for-evaluating-multidialectal-arabic-word-embedding", "name": "DiaLex: A Benchmark for Evaluating Multidialectal Arabic Word Embeddings", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2011.10970" }, "dialects": [ "egy", "lev", "magh", "mixed" ], "year": 2020, "venue": "arXiv 2020", "notes": "We describe DiaLex, a benchmark for intrinsic evaluation of dialectal Arabic word embedding.", "metrics": null }, { "id": "is-this-sentence-valid-an-arabic-dataset-for-commonsense-validation", "name": "Is this sentence valid? An Arabic Dataset for Commonsense Validation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://arxiv.org/abs/2008.10873" }, "year": 2020, "venue": "arXiv 2020", "notes": "We present a benchmark Arabic dataset for commonsense understanding and validation as well as a baseline research and models trained using the same dataset.", "metrics": null }, { "id": "large-arabic-twitter-dataset-on-covid-19", "name": "Large Arabic Twitter Dataset on COVID-19", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "nlp" ], "links": { "paper": "https://arxiv.org/abs/2004.04315" }, "year": 2020, "venue": "arXiv 2020", "notes": "In this work, we describe the first Arabic tweets dataset on COVID-19 that we have been collecting since January 1st, 2020.", "metrics": null }, { "id": "machine-generation-and-detection-of-arabic-manipulated-and-fake-news", "name": "Machine Generation and Detection of Arabic Manipulated and Fake News", "type": "paper", "country": "INTL", "org": "Natural Language Processing Lab", "license": "unknown", "modality": "text", "tasks": [ "fact-checking" ], "links": { "paper": "https://aclanthology.org/2020.wanlp-1.7/" }, "year": 2020, "venue": "WANLP 2020", "notes": "We present a novel method for automatically generating Arabic manipulated (and potentially fake) news stories.", "metrics": null }, { "id": "manorm-a-normalization-dictionary-for-moroccan-arabic-dialect-written", "name": "MANorm: A Normalization Dictionary for Moroccan Arabic Dialect Written in Latin Script", "type": "paper", "country": "MA", "org": "Mohammed V University", "license": "unknown", "modality": "text", "tasks": [ "lexicon" ], "links": { "paper": "https://aclanthology.org/2020.wanlp-1.14/" }, "dialects": [ "magh" ], "year": 2020, "venue": "WANLP 2020", "notes": "This text, however, does not follow the standard rules of writing.", "metrics": null }, { "id": "multi-dialect-arabic-bert-for-country-level-dialect-identification", "name": "Multi-dialect Arabic BERT for Country-level Dialect Identification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id" ], "links": { "paper": "https://aclanthology.org/2020.wanlp-1.10/" }, "year": 2020, "venue": "WANLP 2020", "notes": "We present the experiments conducted, and the models developed by our competing team, Mawdoo3 AI, along the way to achieving our winning solution to subtask.", "metrics": null }, { "id": "offensive-language-detection-in-arabic-using-ulmfit", "name": "Offensive language detection in Arabic using ULMFiT", "type": "paper", "country": "INTL", "org": "Rutgers University - Computer Science", "license": "unknown", "modality": "text", "tasks": [ "offensive-language" ], "links": { "paper": "https://aclanthology.org/2020.osact-1.13/" }, "year": 2020, "venue": "OSACT 2020", "notes": "We approach the shared task OffenseEval 2020 by Mubarak et al.", "metrics": null }, { "id": "osact4-shared-task-on-offensive-language-detection-intensive-preproces", "name": "OSACT4 Shared Task on Offensive Language Detection: Intensive Preprocessing-Based Approach", "type": "paper", "country": "KW", "org": "Kuwait University", "license": "unknown", "modality": "text", "tasks": [ "offensive-language", "classification", "shared-task" ], "links": { "paper": "https://aclanthology.org/2020.osact-1.8/" }, "year": 2020, "venue": "OSACT 2020", "notes": "We apply intensive preprocessing techniques to the dataset before processing it further and feeding it into the classification model.", "metrics": null }, { "id": "parallel-resources-for-tunisian-arabic-dialect-translation", "name": "Parallel resources for Tunisian Arabic Dialect Translation", "type": "paper", "country": "TN", "org": "University of Sfax", "license": "unknown", "modality": "text", "tasks": [ "translation" ], "links": { "paper": "https://aclanthology.org/2020.wanlp-1.18/" }, "dialects": [ "magh" ], "year": 2020, "venue": "WANLP 2020", "notes": "We present a data augmentation technique to create a parallel corpus for Tunisian Arabic dialect written in social media and standard Arabic in order.", "metrics": null }, { "id": "understanding-and-detecting-dangerous-speech-in-social-media", "name": "Understanding and Detecting Dangerous Speech in Social Media", "type": "paper", "country": "INTL", "org": "UBC-NLP", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "offensive-language" ], "links": { "paper": "https://aclanthology.org/2020.osact-1.6/" }, "year": 2020, "venue": "OSACT 2020", "notes": "We report our efforts to build a labeled dataset for dangerous speech.", "metrics": null }, { "id": "woli-at-semeval-2020-task-12-arabic-offensive-language-identification", "name": "WOLI at SemEval-2020 Task 12: Arabic Offensive Language Identification on Different Twitter Datasets", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "nlp" ], "links": { "paper": "https://arxiv.org/abs/2009.05456" }, "year": 2020, "venue": "arXiv 2020", "notes": "This paper presents the results and the main findings of SemEval-2020, Task 12 OffensEval Sub-task A Zampieri et al.", "metrics": null }, { "id": "arabic-natural-language-processing-an-overview", "name": "Arabic natural language processing: An overview", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey" ], "links": { "paper": "https://arxiv.org/abs/1903.02784" }, "year": 2019, "venue": "Journal of King Saud University: Computer and Information Sciences", "citations": 234, "notes": "Arabic is recognised as the 4th most used language of the Internet.", "metrics": null }, { "id": "a-comprehensive-survey-of-arabic-sentiment-analysis", "name": "A comprehensive survey of arabic sentiment analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "sentiment" ], "links": { "paper": "https://doi.org/10.1016/J.IPM.2018.07.006" }, "year": 2019, "venue": "Information Processing & Management", "citations": 201, "notes": "Sentiment analysis (SA) is a continuing field of research that lies at the intersection of many fields such as data mining, natural language processing and", "metrics": null }, { "id": "neural-arabic-question-answering", "name": "Neural Arabic Question Answering", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "qa" ], "links": { "paper": "https://arxiv.org/abs/1906.05394" }, "year": 2019, "venue": "WANLP@ACL 2019", "citations": 177, "notes": "This paper tackles the problem of open domain factual Arabic question answering (QA) using Wikipedia as our knowledge source.", "metrics": null }, { "id": "enhancing-aspect-based-sentiment-analysis-of-arabic-hotels-r", "name": "Enhancing Aspect-Based Sentiment Analysis of Arabic Hotels' reviews using morphological, syntactic and semantic features", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "morphology" ], "links": { "paper": "https://doi.org/10.1016/J.IPM.2018.01.006" }, "year": 2019, "venue": "Information Processing & Management", "citations": 169, "notes": "This research presents an enhanced approach for Aspect-Based Sentiment Analysis (ABSA) of Hotels’ Arabic reviews using supervised machine learning.", "metrics": null }, { "id": "feature-selection-using-binary-grey-wolf-optimizer-with-elit", "name": "Feature selection using binary grey wolf optimizer with elite-based crossover for Arabic text classification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification" ], "links": { "paper": "https://doi.org/10.1007/s00521-019-04368-6" }, "year": 2019, "venue": "Neural computing & applications (Print)", "citations": 167, "metrics": null }, { "id": "the-madar-shared-task-on-arabic-fine-grained-dialect-identif", "name": "The MADAR Shared Task on Arabic Fine-Grained Dialect Identification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id", "evaluation" ], "links": { "paper": "https://aclanthology.org/W19-4622/" }, "year": 2019, "venue": "WANLP@ACL 2019", "citations": 150, "notes": "In this paper, we present the results and findings of the MADAR Shared Task on Arabic Fine-Grained Dialect Identification.", "metrics": null }, { "id": "mazajak-an-online-arabic-sentiment-analyser", "name": "Mazajak: An Online Arabic Sentiment Analyser", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://aclanthology.org/W19-4621/" }, "year": 2019, "venue": "WANLP@ACL 2019", "citations": 141, "notes": "In this paper, we present “Mazajak”, an online system for Arabic SA.", "metrics": null }, { "id": "sanad-single-label-arabic-news-articles-dataset-for-automati", "name": "SANAD: Single-label Arabic News Articles Dataset for automatic text categorization", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification", "dataset" ], "links": { "paper": "https://doi.org/10.1016/j.dib.2019.104076" }, "year": 2019, "venue": "Data in Brief", "citations": 106, "notes": "Text Classification is one of the most popular Natural Language Processing (NLP) tasks.", "metrics": null }, { "id": "deep-learning-approaches-for-arabic-sentiment-analysis", "name": "Deep learning approaches for Arabic sentiment analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1007/s13278-019-0596-4" }, "year": 2019, "venue": "Social Network Analysis and Mining", "citations": 100, "metrics": null }, { "id": "detecting-arabic-depressed-users-from-twitter-data", "name": "Detecting Arabic Depressed Users from Twitter Data", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification" ], "links": { "paper": "https://doi.org/10.1016/j.procs.2019.12.107" }, "year": 2019, "venue": "Procedia Computer Science", "citations": 94, "notes": "Depression is one of the most common health issues impacting the world.", "metrics": null }, { "id": "idat-at-fire2019-overview-of-the-track-on-irony-detection-in", "name": "IDAT at FIRE2019: Overview of the Track on Irony Detection in Arabic Tweets", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "classification" ], "links": { "paper": "https://doi.org/10.1145/3368567.3368585" }, "year": 2019, "venue": "Fire", "citations": 92, "notes": "This overview paper describes the first shared task on irony detection for the Arabic language.", "metrics": null }, { "id": "a-survey-of-opinion-mining-in-arabic", "name": "A Survey of Opinion Mining in Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey", "sentiment" ], "links": { "paper": "https://doi.org/10.1145/3295662" }, "year": 2019, "venue": "ACM Trans. Asian Low Resour. Lang. Inf. Process.", "citations": 86, "notes": "Opinion-mining or sentiment analysis continues to gain interest in industry and academics.", "metrics": null }, { "id": "detecting-offensive-language-on-arabic-social-media-using-de", "name": "Detecting Offensive Language on Arabic Social Media Using Deep Learning", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification" ], "links": { "paper": "https://doi.org/10.1109/SNAMS.2019.8931839" }, "year": 2019, "venue": "International Conference on Social Networks Analysis, Management and Security", "citations": 86, "notes": "Offensive content on social media such as verbal attacks, demeaning comments or hate speech has many negative effects on its users.", "metrics": null }, { "id": "arabic-natural-language-processing-and-machine-learning-base", "name": "Arabic Natural Language Processing and Machine Learning-Based Systems", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2018.2890076" }, "year": 2019, "venue": "IEEE Access", "citations": 84, "notes": "Arabic natural language processing (ANLP) consists of developing techniques and tools that can utilize and analyze the Arabic language in both written and", "metrics": null }, { "id": "hilatsa-a-hybrid-incremental-learning-approach-for-arabic-tw", "name": "HILATSA: A hybrid Incremental learning approach for Arabic tweets sentiment analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1016/J.EIJ.2019.03.002" }, "year": 2019, "venue": "Egyptian Informatics Journal", "citations": 80, "notes": "A huge amount of data is generated since the evolution in technology and the tremendous growth of social networks.", "metrics": null }, { "id": "the-mgb-5-challenge-recognition-and-dialect-identification-o", "name": "The MGB-5 Challenge: Recognition and Dialect Identification of Dialectal Arabic Speech", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "dialect-id" ], "links": { "paper": "https://doi.org/10.1109/ASRU46091.2019.9003960" }, "year": 2019, "venue": "Automatic Speech Recognition & Understanding", "citations": 80, "notes": "This paper describes the fifth edition of the Multi-Genre Broadcast Challenge (MGB-5), an evaluation focused on Arabic speech recognition and dialect", "metrics": null }, { "id": "arsentd-lev-a-multi-topic-corpus-for-target-based-sentiment", "name": "ArSentD-LEV: A Multi-Topic Corpus for Target-based Sentiment Analysis in Arabic Levantine Tweets", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "dataset" ], "links": { "paper": "https://arxiv.org/abs/1906.01830" }, "year": 2019, "venue": "arXiv.org", "citations": 78, "notes": "Sentiment analysis is a highly subjective and challenging task.", "metrics": null }, { "id": "a-study-of-the-effects-of-stemming-strategies-on-arabic-docu", "name": "A Study of the Effects of Stemming Strategies on Arabic Document Classification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "preprocessing", "classification" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2019.2903331" }, "year": 2019, "venue": "IEEE Access", "citations": 70, "notes": "Stemming is one of the most effective techniques, which has been adopted in many applications, such as machine learning, machine translation, document", "metrics": null }, { "id": "surface-and-deep-features-ensemble-for-sentiment-analysis-of", "name": "Surface and Deep Features Ensemble for Sentiment Analysis of Arabic Tweets", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2019.2924314" }, "year": 2019, "venue": "IEEE Access", "citations": 70, "notes": "Sentiment analysis (SA) of Arabic tweets is a complex task due to the rich morphology of the Arabic language and the informal nature of language on Twitter.", "metrics": null }, { "id": "an-efficient-single-document-arabic-text-summarization-using", "name": "An efficient single document Arabic text summarization using a combination of statistical and semantic features", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "summarization" ], "links": { "paper": "https://doi.org/10.1016/j.jksuci.2019.03.010" }, "year": 2019, "venue": "Journal of King Saud University: Computer and Information Sciences", "citations": 67, "notes": "In this paper, we propose an automatic, generic, and extractive Arabic single", "metrics": null }, { "id": "hulmona-the-universal-language-model-in-arabic", "name": "hULMonA: The Universal Language Model in Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://aclanthology.org/W19-4608/" }, "year": 2019, "venue": "WANLP@ACL 2019", "citations": 64, "notes": "Arabic is a complex language with limited resources which makes it challenging to produce accurate text classification tasks such as sentiment analysis.", "metrics": null }, { "id": "highly-effective-arabic-diacritization-using-sequence-to-seq", "name": "Highly Effective Arabic Diacritization using Sequence to Sequence Modeling", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "diacritization" ], "links": { "paper": "https://aclanthology.org/N19-1248/" }, "year": 2019, "venue": "North American Chapter of the Association for Computational Linguistics", "citations": 51, "notes": "Arabic text is typically written without short vowels (or diacritics).", "metrics": null }, { "id": "arhnet-leveraging-community-interaction-for-detection-of-rel", "name": "ARHNet - Leveraging Community Interaction for Detection of Religious Hate Speech in Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification" ], "links": { "paper": "https://aclanthology.org/P19-2038/" }, "year": 2019, "venue": "Annual Meeting of the Association for Computational Linguistics", "citations": 35, "notes": "The rapid widespread of social media has lead to some undesirable consequences like the rapid increase of hateful content and offensive language.", "metrics": null }, { "id": "no-army-no-navy-bert-semi-supervised-learning-of-arabic-dial", "name": "No Army, No Navy: BERT Semi-Supervised Learning of Arabic Dialects", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "paper": "https://aclanthology.org/W19-4637/" }, "year": 2019, "venue": "WANLP@ACL 2019", "citations": 34, "notes": "We present our deep leaning system submitted to MADAR shared task 2 focused on twitter user dialect identification.", "metrics": null }, { "id": "adida-automatic-dialect-identification-for-arabic", "name": "ADIDA: Automatic Dialect Identification for Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id" ], "links": { "paper": "https://aclanthology.org/N19-4002/" }, "year": 2019, "venue": "North American Chapter of the Association for Computational Linguistics", "citations": 26, "notes": "This demo paper describes ADIDA, a web-based system for automatic dialect identification for Arabic text.", "metrics": null }, { "id": "morphologically-annotated-corpora-for-seven-arabic-dialects", "name": "Morphologically Annotated Corpora for Seven Arabic Dialects: Taizi, Sanaani, Najdi, Jordanian, Syrian, Iraqi and Moroccan", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "morphology", "dataset" ], "links": { "paper": "https://aclanthology.org/W19-4615/" }, "year": 2019, "venue": "WANLP@ACL 2019", "citations": 25, "notes": "We present a collection of morphologically annotated corpora for seven Arabic dialects: Taizi Yemeni, Sanaani Yemeni, Najdi, Jordanian, Syrian, Iraqi and", "metrics": null }, { "id": "arabic-named-entity-recognition-what-works-and-what-s-next", "name": "Arabic Named Entity Recognition: What Works and What’s Next", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "ner" ], "links": { "paper": "https://aclanthology.org/W19-4607/" }, "year": 2019, "venue": "WANLP@ACL 2019", "citations": 23, "notes": "This paper presents the winning solution to the Arabic Named Entity Recognition challenge run by Topcoder.com.", "metrics": null }, { "id": "arbengvec-arabic-english-cross-lingual-word-embedding-model", "name": "ArbEngVec : Arabic-English Cross-Lingual Word Embedding Model", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding" ], "links": { "paper": "https://aclanthology.org/W19-4605/" }, "year": 2019, "venue": "WANLP@ACL 2019", "citations": 19, "notes": "Word Embeddings (WE) are getting increasingly popular and widely applied in many Natural Language Processing (NLP) applications due to their effectiveness in", "metrics": null }, { "id": "arabic-dialect-identification-for-travel-and-twitter-text", "name": "Arabic Dialect Identification for Travel and Twitter Text", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id" ], "links": { "paper": "https://aclanthology.org/W19-4628/" }, "year": 2019, "venue": "WANLP@ACL 2019", "citations": 17, "notes": "This paper presents the results of the experiments done as a part of MADAR Shared Task in WANLP 2019 on Arabic Fine-Grained Dialect Identification.", "metrics": null }, { "id": "arabic-rule-based-named-entity-recognition-system-using-gate", "name": "Arabic Rule-based Named Entity Recognition System Using GATE", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "ner" ], "links": { "paper": "https://www.semanticscholar.org/paper/791020059cb28d5582104fa4e90478ba88c03f5f" }, "year": 2019, "venue": "IAPR International Conference on Machine Learning and Data Mining in Pattern Recognition", "citations": 17, "metrics": null }, { "id": "arabic-tweet-act-speech-act-recognition-for-arabic-asynchron", "name": "Arabic Tweet-Act: Speech Act Recognition for Arabic Asynchronous Conversations", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dataset" ], "links": { "paper": "https://aclanthology.org/W19-4620/" }, "year": 2019, "venue": "WANLP@ACL 2019", "citations": 16, "notes": "In this paper, we proposed speech act classification for asynchronous conversations on Twitter using multiple machine learning methods including SVM and deep", "metrics": null }, { "id": "mawdoo3-ai-at-madar-shared-task-arabic-fine-grained-dialect", "name": "Mawdoo3 AI at MADAR Shared Task: Arabic Fine-Grained Dialect Identification with Ensemble Learning", "type": "paper", "country": "JO", "org": "Mawdoo3", "license": "unknown", "modality": "text", "tasks": [ "dialect-id", "evaluation" ], "links": { "paper": "https://aclanthology.org/W19-4630/" }, "year": 2019, "venue": "WANLP@ACL 2019", "citations": 16, "notes": "In this paper we discuss several models we used to classify 25 city-level Arabic dialects in addition to Modern Standard Arabic (MSA) as part of MADAR shared", "metrics": null }, { "id": "st-madar-2019-shared-task-arabic-fine-grained-dialect-identi", "name": "ST MADAR 2019 Shared Task: Arabic Fine-Grained Dialect Identification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id", "evaluation" ], "links": { "paper": "https://aclanthology.org/W19-4635/" }, "year": 2019, "venue": "WANLP@ACL 2019", "citations": 16, "notes": "This paper describes the solution that we propose on MADAR 2019 Arabic Fine-Grained Dialect Identification task.", "metrics": null }, { "id": "the-madar-arabic-dialect-corpus-and-lexicon", "name": "The MADAR Arabic Dialect Corpus and Lexicon", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dataset" ], "links": { "paper": "https://aclanthology.org/L18-1535/" }, "year": 2018, "venue": "International Conference on Language Resources and Evaluation", "citations": 324, "notes": "In this paper, we present two resources that were created as part of the Multi Arabic Dialect Applications and Resources (MADAR) project.", "metrics": null }, { "id": "are-they-our-brothers-analysis-and-detection-of-religious-ha", "name": "Are They Our Brothers? Analysis and Detection of Religious Hate Speech in the Arabic Twittersphere", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification" ], "links": { "paper": "https://www.semanticscholar.org/paper/70ab0898e5325bf33330e96668a5771ef35d32d3" }, "year": 2018, "venue": "arXiv 2018", "citations": 251, "metrics": null }, { "id": "improved-whale-optimization-algorithm-for-feature-selection", "name": "Improved whale optimization algorithm for feature selection in Arabic sentiment analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1007/s10489-018-1334-8" }, "year": 2018, "venue": "Applied intelligence (Boston)", "citations": 251, "metrics": null }, { "id": "sentiment-analysis-of-arabic-tweets-using-deep-learning", "name": "Sentiment Analysis of Arabic Tweets using Deep Learning", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1016/J.PROCS.2018.10.466" }, "year": 2018, "venue": "International Conference on Arabic Computational Linguistics", "citations": 245, "notes": "Sentiment analysis is the computational study of people’s opinions, attitudes and emotions toward entities, individuals, issues, events or topics.", "metrics": null }, { "id": "using-long-short-term-memory-deep-neural-networks-for-aspect", "name": "Using long short-term memory deep neural networks for aspect-based sentiment analysis of Arabic reviews", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1007/s13042-018-0799-4" }, "year": 2018, "venue": "International Journal of Machine Learning and Cybernetics", "citations": 245, "metrics": null }, { "id": "a-combined-cnn-and-lstm-model-for-arabic-sentiment-analysis", "name": "A Combined CNN and LSTM Model for Arabic Sentiment Analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://arxiv.org/abs/1807.02911" }, "year": 2018, "venue": "International Cross-Domain Conference on Machine Learning and Knowledge Extraction", "citations": 170, "notes": "Deep neural networks have shown good data modelling capabilities when dealing with challenging and large datasets from a wide range of application areas.", "metrics": null }, { "id": "hotel-arabic-reviews-dataset-construction-for-sentiment-anal", "name": "Hotel Arabic-Reviews Dataset Construction for Sentiment Analysis Applications", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "dataset" ], "links": { "paper": "https://doi.org/10.1007/978-3-319-67056-0_3" }, "year": 2018, "venue": "arXiv 2018", "citations": 161, "metrics": null }, { "id": "fine-grained-arabic-dialect-identification", "name": "Fine-Grained Arabic Dialect Identification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id" ], "links": { "paper": "https://aclanthology.org/C18-1113/" }, "year": 2018, "venue": "International Conference on Computational Linguistics", "citations": 139, "notes": "This paper presents the first results on a fine-grained dialect classification task covering 25 specific cities from across the Arab World, in addition to", "metrics": null }, { "id": "towards-accurate-detection-of-offensive-language-in-online-c", "name": "Towards Accurate Detection of Offensive Language in Online Communication in Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "hate-speech", "classification" ], "links": { "paper": "https://doi.org/10.1016/j.procs.2018.10.491" }, "year": 2018, "venue": "International Conference on Arabic Computational Linguistics", "citations": 117, "notes": "We present the results of predictive modelling for the detection of anti-social behaviour in online communication in Arabic, such as comments which contain", "metrics": null }, { "id": "automatic-arabic-dialect-classification-using-deep-learning", "name": "Automatic Arabic Dialect Classification Using Deep Learning Models", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id", "classification" ], "links": { "paper": "https://doi.org/10.1016/J.PROCS.2018.10.489" }, "year": 2018, "venue": "International Conference on Arabic Computational Linguistics", "citations": 112, "notes": "Recently, the vast use of social media and the high availability of internet access have produced a considerably different textual data from the formal and", "metrics": null }, { "id": "dataset-construction-for-the-detection-of-anti-social-behavi", "name": "Dataset Construction for the Detection of Anti-Social Behaviour in Online Communication in Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification", "dataset" ], "links": { "paper": "https://doi.org/10.1016/J.PROCS.2018.10.473" }, "year": 2018, "venue": "International Conference on Arabic Computational Linguistics", "citations": 112, "notes": "Warning: this paper contains a range of words which may cause offence.", "metrics": null }, { "id": "arsas-an-arabic-speech-act-and-sentiment-corpus-of-tweets", "name": "ArSAS : An Arabic Speech-Act and Sentiment Corpus of Tweets", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "dataset" ], "links": { "paper": "https://www.semanticscholar.org/paper/d32d3bb226f1738f72c415c6b03b5ad66ff604a4" }, "year": 2018, "venue": "arXiv 2018", "citations": 100, "metrics": null }, { "id": "unified-guidelines-and-resources-for-arabic-dialect-orthogra", "name": "Unified Guidelines and Resources for Arabic Dialect Orthography", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "morphology", "dataset" ], "links": { "paper": "https://aclanthology.org/L18-1574/" }, "year": 2018, "venue": "International Conference on Language Resources and Evaluation", "citations": 96, "metrics": null }, { "id": "sedat-sentiment-and-emotion-detection-in-arabic-text-using-c", "name": "SEDAT: Sentiment and Emotion Detection in Arabic Text Using CNN-LSTM Deep Learning", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "classification" ], "links": { "paper": "https://doi.org/10.1109/ICMLA.2018.00134" }, "year": 2018, "venue": "International Conference on Machine Learning and Applications", "citations": 94, "notes": "Social media is growing as a communication medium where people can express online their feelings and opinions on a variety of topics in ways they rarely do in", "metrics": null }, { "id": "a-survey-of-arabic-text-mining", "name": "A Survey of Arabic Text Mining", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "survey" ], "links": { "paper": "https://doi.org/10.1007/978-3-319-67056-0_20" }, "year": 2018, "venue": "arXiv 2018", "citations": 89, "metrics": null }, { "id": "improving-sentiment-analysis-in-arabic-using-word-representa", "name": "Improving Sentiment Analysis in Arabic Using Word Representation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "embedding" ], "links": { "paper": "https://arxiv.org/abs/1803.00124" }, "year": 2018, "venue": "International Workshop on Arabic Script Analysis and Recognition", "citations": 88, "notes": "The complexities of Arabic language in morphology, orthography and dialects makes sentiment analysis for Arabic more challenging.", "metrics": null }, { "id": "shami-a-corpus-of-levantine-arabic-dialects", "name": "Shami: A Corpus of Levantine Arabic Dialects", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dataset" ], "links": { "paper": "https://aclanthology.org/L18-1576/" }, "year": 2018, "venue": "International Conference on Language Resources and Evaluation", "citations": 87, "notes": "Modern Standard Arabic (MSA) is the official language used in education and media across the Arab world both in writing and formal speech.", "metrics": null }, { "id": "revisiting-k-means-and-topic-modeling-a-comparison-study-to", "name": "Revisiting K-Means and Topic Modeling, a Comparison Study to Cluster Arabic Documents", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "classification" ], "links": { "paper": "https://doi.org/10.1109/ACCESS.2018.2852648" }, "year": 2018, "venue": "IEEE Access", "citations": 86, "notes": "This paper uses a combined method to cluster Arabic text documents.", "metrics": null }, { "id": "you-tweet-what-you-speak-a-city-level-dataset-of-arabic-dial", "name": "You Tweet What You Speak: A City-Level Dataset of Arabic Dialects", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dataset" ], "links": { "paper": "https://aclanthology.org/L18-1577/" }, "year": 2018, "venue": "International Conference on Language Resources and Evaluation", "citations": 82, "notes": "Arabic has a wide range of varieties or dialects.", "metrics": null }, { "id": "a-lexical-distance-study-of-arabic-dialects", "name": "A Lexical Distance Study of Arabic Dialects", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "paper": "https://doi.org/10.1016/J.PROCS.2018.10.456" }, "year": 2018, "venue": "International Conference on Arabic Computational Linguistics", "citations": 80, "notes": "Diglossia is a very common phenomenon in Arabic-speaking communities, where the spoken language is different from both Classical Arabic (CA) and Modern Standard", "metrics": null }, { "id": "a-hybrid-approach-for-arabic-text-summarization-using-domain", "name": "A Hybrid Approach for Arabic Text Summarization Using Domain Knowledge and Genetic Algorithms", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "summarization" ], "links": { "paper": "https://doi.org/10.1007/s12559-018-9547-z" }, "year": 2018, "venue": "Cognitive Computation", "citations": 70, "metrics": null }, { "id": "deep-models-for-arabic-dialect-identification-on-benchmarked", "name": "Deep Models for Arabic Dialect Identification on Benchmarked Data", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id", "evaluation" ], "links": { "paper": "https://aclanthology.org/W18-3930/" }, "year": 2018, "venue": "VarDial@COLING 2018", "citations": 69, "notes": "We treat these two limitations:We (1) benchmark the data, and (2) empirically test6different deep learning methods on thetask, comparing peformance to several", "metrics": null }, { "id": "automatic-arabic-image-captioning-using-rnn-lstm-based-langu", "name": "Automatic Arabic Image Captioning using RNN-LSTM-Based Language Model and CNN", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "multimodal", "tasks": [ "multimodal", "pretraining" ], "links": { "paper": "https://doi.org/10.14569/IJACSA.2018.090610" }, "year": 2018, "venue": "arXiv 2018", "citations": 64, "notes": "The automatic generation of correct syntaxial and semantical image captions is an essential problem in Artificial Intelligence.", "metrics": null }, { "id": "a-neural-machine-translation-model-for-arabic-dialects-that", "name": "A Neural Machine Translation Model for Arabic Dialects That Utilises Multitask Learning (MTL)", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "translation" ], "links": { "paper": "https://doi.org/10.1155/2018/7534712" }, "year": 2018, "venue": "Computational Intelligence and Neuroscience", "citations": 63, "notes": "In this research article, we study the problem of employing a neural machine translation model to translate Arabic dialects to modern standard Arabic.", "metrics": null }, { "id": "an-abstractive-arabic-text-summarizer-with-user-controlled-g", "name": "An abstractive Arabic text summarizer with user controlled granularity", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "summarization" ], "links": { "paper": "https://doi.org/10.1016/J.IPM.2018.06.002" }, "year": 2018, "venue": "Information Processing & Management", "citations": 63, "notes": "In the former we retain the more important sentences more or less in their original structure, while the latter requires a fusion of multiple", "metrics": null }, { "id": "image-based-arabic-sign-language-recognition-system", "name": "Image based Arabic Sign Language Recognition System", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "vision", "tasks": [ "multimodal" ], "links": { "paper": "https://doi.org/10.14569/IJACSA.2018.090327" }, "year": 2018, "venue": "arXiv 2018", "citations": 62, "notes": "Through history, humans have used many ways of communication such as gesturing, sounds, drawing, writing, and speaking.", "metrics": null }, { "id": "sentialg-automated-corpus-annotation-for-algerian-sentiment", "name": "SentiALG: Automated Corpus Annotation for Algerian Sentiment Analysis", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "dataset" ], "links": { "paper": "https://arxiv.org/abs/1808.05079" }, "year": 2018, "venue": "International Conference on Advances in Brain Inspired Cognitive Systems", "citations": 61, "notes": "To sort a text into two classes, the very first thing we need is a good annotation guideline, establishing what is required to qualify for each class.", "metrics": null }, { "id": "sentiment-lexicon-for-sentiment-analysis-of-saudi-dialect-tw", "name": "Sentiment lexicon for sentiment analysis of Saudi dialect tweets", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "dataset" ], "links": { "paper": "https://doi.org/10.1016/J.PROCS.2018.10.494" }, "year": 2018, "venue": "International Conference on Arabic Computational Linguistics", "citations": 61, "notes": "Twitter is one of the most widely used social media platforms in Saudi Arabia and is a rich source for mining the public’s attitude towards political, social", "metrics": null }, { "id": "a-morphologically-annotated-corpus-of-emirati-arabic", "name": "A Morphologically Annotated Corpus of Emirati Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "morphology", "dataset" ], "links": { "paper": "https://aclanthology.org/L18-1607/" }, "year": 2018, "venue": "International Conference on Language Resources and Evaluation", "citations": 55, "metrics": null }, { "id": "comparison-of-pre-trained-word-vectors-for-arabic-text-class", "name": "Comparison of Pre-Trained Word Vectors for Arabic Text Classification Using Deep Learning Approach", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "embedding", "classification", "pretraining" ], "links": { "paper": "https://doi.org/10.1109/ICMLA.2018.00239" }, "year": 2018, "venue": "International Conference on Machine Learning and Applications", "citations": 49, "notes": "In this paper, we describe an Arabic text sentiment analysis approach using a Deep Neural network,", "metrics": null }, { "id": "multi-dialect-arabic-pos-tagging-a-crf-approach", "name": "Multi-Dialect Arabic POS Tagging: A CRF Approach", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "morphology" ], "links": { "paper": "https://aclanthology.org/L18-1015/" }, "year": 2018, "venue": "International Conference on Language Resources and Evaluation", "citations": 45, "notes": "This paper introduces a new dataset of POS-tagged Arabic tweets in four major dialects along with tagging guidelines.", "metrics": null }, { "id": "arabic-dialect-identification-in-the-context-of-bivalency-an", "name": "Arabic Dialect Identification in the Context of Bivalency and Code-Switching", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id" ], "links": { "paper": "https://aclanthology.org/L18-1573/" }, "year": 2018, "venue": "International Conference on Language Resources and Evaluation", "citations": 41, "notes": "In this paper we use a novel approach towards Arabic dialect identification using language bivalency and written code-switching.", "metrics": null }, { "id": "collection-and-analysis-of-code-switch-egyptian-arabic-engli", "name": "Collection and Analysis of Code-switch Egyptian Arabic-English Speech Corpus", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "speech", "tasks": [ "dataset" ], "links": { "paper": "https://aclanthology.org/L18-1601/" }, "year": 2018, "venue": "International Conference on Language Resources and Evaluation", "citations": 37, "notes": "Speech corpora are key components needed by both: linguists (in language analyses, research and teaching languages) and Natural Language Processing (NLP)", "metrics": null }, { "id": "character-level-convolutional-neural-network-for-arabic-dial", "name": "Character Level Convolutional Neural Network for Arabic Dialect Identification", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id" ], "links": { "paper": "https://aclanthology.org/W18-3913/" }, "year": 2018, "venue": "VarDial@COLING 2018", "citations": 35, "metrics": null }, { "id": "part-of-speech-tagging-for-arabic-gulf-dialect-using-bi-lstm", "name": "Part-of-Speech Tagging for Arabic Gulf Dialect Using Bi-LSTM", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "morphology" ], "links": { "paper": "https://aclanthology.org/L18-1620/" }, "year": 2018, "venue": "International Conference on Language Resources and Evaluation", "citations": 31, "notes": "Part-of-speech (POS) tagging is one of the most important addressed areas in the natural language processing (NLP).", "metrics": null }, { "id": "arabizi-sentiment-analysis-based-on-transliteration-and-auto", "name": "Arabizi sentiment analysis based on transliteration and automatic corpus annotation", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "preprocessing", "dataset" ], "links": { "paper": "https://aclanthology.org/W18-6249/" }, "year": 2018, "venue": "WASSA@EMNLP", "citations": 26, "notes": "Arabizi is a form of writing Arabic text which relies on Latin letters, numerals and punctuation rather than Arabic letters.", "metrics": null }, { "id": "noise-robust-morphological-disambiguation-for-dialectal-arab", "name": "Noise-Robust Morphological Disambiguation for Dialectal Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "morphology" ], "links": { "paper": "https://aclanthology.org/N18-1087/" }, "year": 2018, "venue": "North American Chapter of the Association for Computational Linguistics", "citations": 25, "notes": "User-generated text tends to be noisy with many lexical and orthographic inconsistencies, making natural language processing (NLP) tasks more challenging.", "metrics": null }, { "id": "unibuckernel-reloaded-first-place-in-arabic-dialect-identifi", "name": "UnibucKernel Reloaded: First Place in Arabic Dialect Identification for the Second Year in a Row", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dialect-id" ], "links": { "paper": "https://arxiv.org/abs/1805.04876" }, "year": 2018, "venue": "VarDial@COLING 2018", "citations": 24, "notes": "We present a machine learning approach that ranked on the first place in the Arabic Dialect Identification (ADI) Closed Shared Tasks of the 2018 VarDial", "metrics": null }, { "id": "a-leveled-reading-corpus-of-modern-standard-arabic", "name": "A Leveled Reading Corpus of Modern Standard Arabic", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "tasks": [ "dataset" ], "links": { "paper": "https://aclanthology.org/L18-1366/" }, "year": 2018, "venue": "International Conference on Language Resources and Evaluation", "citations": 23, "notes": "We present a reading corpus in Modern Standard Arabic to enrich the sparse collection of resources that can be leveraged for educational applications.", "metrics": null }, { "id": "deep-learning-for-arabic-nlp-a-survey", "name": "Deep Learning for Arabic NLP: A Survey", "type": "paper", "country": "INTL", "org": "Various", "license": "unknown", "modality": "text", "year": 2018, "venue": "Journal of Computational Science 2018", "tasks": [ "survey", "nlp" ], "links": { "paper": "https://www.sciencedirect.com/science/article/pii/S1877750317303757" }, "notes": "Survey of deep learning methods across Arabic NLP tasks.", "metrics": null }, { "id": "jais-family-590m", "name": "Jais-family-590M", "type": "llm", "country": "AE", "org": "Inception AI (G42)", "license": "apache-2.0", "modality": "text", "tasks": [ "pretraining", "chat" ], "links": { "hf": "https://huggingface.co/inception42/jais-family-590m" }, "base_model": [ "from-scratch" ], "size": "590M", "year": 2024, "on_device": true, "notes": "Smallest Jais-family Arabic-English model, suitable for edge and on-device use.", "metrics": { "downloads": 1985, "likes": 7, "lastModified": "2024-09-11" } }, { "id": "aramodernbert-base-v1-0", "name": "AraModernBert-Base-V1.0", "type": "llm", "country": "SA", "org": "NAMAA-Space", "license": "apache-2.0", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/NAMAA-Space/AraModernBert-Base-V1.0" }, "size": "150M", "year": 2025, "on_device": true, "dialects": [ "msa" ], "notes": "ModernBERT-style Arabic encoder with long context from NAMAA.", "metrics": { "downloads": 1976, "likes": 15, "lastModified": "2026-03-30" } }, { "id": "arabic-english-handwritten-ocr-v3", "name": "Arabic-English-handwritten-OCR-v3", "type": "ocr", "country": "INTL", "org": "sherif1313", "license": "apache-2.0", "modality": "vision", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/sherif1313/Arabic-English-handwritten-OCR-v3" }, "notes": "Qwen2.5-VL-3B, trained on 47K handwriting samples", "base_model": [ "qwen/qwen2.5-vl-3b-instruct" ], "metrics": { "downloads": 1972, "likes": 14, "lastModified": "2025-12-29" } }, { "id": "qwencleo-asr", "name": "QwenCleo-ASR", "type": "asr", "country": "EG", "org": "Mohammed Aly", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/mohammedaly22/QwenCleo-ASR" }, "dialects": [ "egy" ], "year": 2026, "notes": "Arabic automatic speech recognition model fine-tuned from Qwen/Qwen3-ASR-1.7B.", "base_model": [ "qwen/qwen3-asr-1.7b" ], "metrics": { "downloads": 1955, "likes": 10, "lastModified": "2026-06-17" } }, { "id": "synth-shamela-ocr-arabic-books", "name": "synth shamela ocr arabic books", "type": "dataset", "country": "INTL", "org": "freococo", "license": "cc-by-nc-nd-4.0", "modality": "multimodal", "tasks": [ "ocr", "vision" ], "links": { "hf": "https://huggingface.co/datasets/freococo/synth_shamela_ocr_arabic_books" }, "size": "1M–10M rows", "year": 2026, "notes": "Structured book pages rendered dynamically with style, font, and degradation variations.", "metrics": { "downloads": 1945, "likes": 2, "lastModified": "2026-07-13" } }, { "id": "whisper-large-v3-turbo-darija", "name": "whisper-large-v3-turbo-darija", "type": "asr", "country": "MA", "org": "Anas Zil", "license": "mit", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/anaszil/whisper-large-v3-turbo-darija" }, "year": 2025, "dialects": [ "magh" ], "notes": "Whisper large-v3-turbo fine-tuned on Moroccan Darija.", "base_model": [ "openai/whisper-large-v3-turbo" ], "metrics": { "downloads": 1927, "likes": 9, "lastModified": "2025-11-09" } }, { "id": "silma-1-0", "name": "SILMA 1.0", "type": "llm", "country": "INTL", "org": "SILMA AI", "license": "gemma", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/silma-ai/SILMA-9B-Instruct-v1.0" }, "size": "9B", "dialects": [ "msa" ], "notes": "Top-ranked Arabic LLM built on Gemma", "metrics": { "downloads": 1912, "likes": 89, "lastModified": "2025-06-25" } }, { "id": "dziribert", "name": "DziriBERT", "type": "llm", "country": "DZ", "org": "alger-ia", "license": "apache-2.0", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/alger-ia/dziribert" }, "size": "124M", "year": 2021, "dialects": [ "magh" ], "notes": "BERT for Algerian dialect incl. Arabizi (Algeria); first Algerian language model.", "metrics": { "downloads": 1909, "likes": 31, "lastModified": "2023-03-17" } }, { "id": "mawrooth-allam-7b-lora", "name": "Mawrooth-ALLaM-7B-LoRA", "type": "llm", "country": "INTL", "org": "NorahAlobaied", "license": "apache-2.0", "modality": "text", "tasks": [ "poetry", "explanation" ], "links": { "hf": "https://huggingface.co/NorahAlobaied/Mawrooth-ALLaM-7B-LoRA" }, "base_model": [ "humain-ai/allam-7b-instruct-preview" ], "dialects": [ "gulf" ], "size": "7B", "on_device": false, "year": 2026, "notes": "ALLaM-7B LoRA adapter that explains Nabati poetry verses in five structured sections.", "metrics": { "downloads": 1909, "likes": 0, "lastModified": "2026-09-27" } }, { "id": "warsh-segments-v3", "name": "warsh segments v3", "type": "dataset", "country": "INTL", "org": "Haitam03", "license": "cc-by-nc-4.0", "modality": "speech", "tasks": [ "asr", "quran" ], "links": { "hf": "https://huggingface.co/datasets/Haitam03/warsh-segments-v3" }, "dialects": [ "classical" ], "size": "100K–1M rows", "year": 2026, "notes": "Warsh (Rewayat Warsh A'n Nafi') Quran recitation, segmented at waqf with obadx/recitation-segmenter-v2.", "metrics": { "downloads": 1906, "likes": 0, "lastModified": "2026-08-28" } }, { "id": "alghafa-native", "name": "AlGhafa Arabic LLM Benchmark (Native)", "type": "benchmark", "country": "INTL", "org": "Open Arabic LLM Leaderboard (OALL)", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/OALL/AlGhafa-Arabic-LLM-Benchmark-Native" }, "year": 2024, "dialects": [ "msa" ], "notes": "Native-Arabic multiple-choice evaluation tasks used as OALL leaderboard tasks.", "metrics": { "downloads": 1846, "likes": 8, "lastModified": "2024-03-07" } }, { "id": "jais-adapted", "name": "Jais-adapted (Llama-2 based)", "type": "llm", "country": "AE", "org": "Inception AI (G42)", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/inception42/jais-adapted-13b-chat" }, "base_model": [ "inceptionai/jais-adapted-13b" ], "size": "7B-70B", "year": 2024, "notes": "Jais-adapted series (7B, 13B, 70B) continued-pretrained from Llama-2 with Arabic vocabulary extension.", "metrics": { "downloads": 1825, "likes": 7, "lastModified": "2024-09-11" } }, { "id": "recitation-segmenter-v2", "name": "recitation segmenter v2", "type": "asr", "country": "INTL", "org": "obadx", "license": "mit", "modality": "speech", "tasks": [ "asr", "evaluation", "quran" ], "links": { "hf": "https://huggingface.co/obadx/recitation-segmenter-v2" }, "dialects": [ "classical" ], "year": 2025, "notes": "This model is a fine-tuned version of facebook/w2v-bert-2.0 for segmenting Holy Quran recitations based on pause points (waqf).", "base_model": [ "facebook/w2v-bert-2.0" ], "metrics": { "downloads": 1777, "likes": 3, "lastModified": "2025-09-03" } }, { "id": "fastconformer-quran", "name": "fastconformer-quran", "type": "asr", "country": "INTL", "org": "Muno459", "license": "other", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/Muno459/fastconformer-quran" }, "year": 2026, "dialects": [ "classical" ], "notes": "FastConformer CTC model for Quran recitation.", "base_model": [ "nvidia/stt_ar_fastconformer_hybrid_large_pcd_v1.0" ], "metrics": { "downloads": 1725, "likes": 26, "lastModified": "2026-08-15" } }, { "id": "bert-large-arabertv2", "name": "bert-large-arabertv2", "type": "llm", "country": "LB", "org": "AUB MIND Lab", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/aubmindlab/bert-large-arabertv2" }, "size": "371M", "year": 2022, "on_device": true, "dialects": [ "msa" ], "notes": "AraBERT v2 large with Farasa-segmented input.", "metrics": { "downloads": 1700, "likes": 12, "lastModified": "2023-03-20" } }, { "id": "bert-base-qarib", "name": "bert base qarib", "type": "llm", "country": "QA", "org": "QCRI", "license": "unknown", "modality": "text", "tasks": [ "encoder" ], "links": { "hf": "https://huggingface.co/ahmedabdelali/bert-base-qarib" }, "year": 2021, "notes": "QARiB: QCRI Arabic and Dialectal BERT trained on about 420M tweets and 180M sentences of text.", "dialects": [ "mixed" ], "metrics": { "downloads": 1697, "likes": 9, "lastModified": "2021-05-20" } }, { "id": "command-r7b-arabic", "name": "Command R7B Arabic", "type": "llm", "country": "INTL", "org": "Cohere", "license": "cc-by-nc-4.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/CohereLabs/c4ai-command-r7b-arabic-02-2025" }, "base_model": [ "coherelabs/c4ai-command-r7b-12-2024" ], "size": "7B", "notes": "Arabic-optimized Command R variant", "metrics": { "downloads": 1673, "likes": 132, "lastModified": "2025-10-30" } }, { "id": "arabic-emirati-female-piper", "name": "arabic-emirati-female-piper", "type": "tts", "country": "AE", "org": "Vadim Belsky", "license": "mit", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/vadimbelsky/arabic-emirati-female-piper" }, "year": 2025, "dialects": [ "gulf" ], "notes": "Piper TTS voice for Emirati Arabic, female speaker.", "metrics": { "downloads": 1665, "likes": 1, "lastModified": "2025-12-01" } }, { "id": "adi17", "name": "ADI17", "type": "dataset", "country": "INTL", "org": "Elyadata", "license": "unknown", "modality": "speech", "tasks": [ "dialect-id" ], "links": { "hf": "https://huggingface.co/datasets/ArabicSpeech/ADI17" }, "size": "3000h", "year": 2020, "dialects": [ "mixed" ], "notes": "Arabic dialect identification speech dataset from YouTube covering 17 countries (Tunisia/Elyadata).", "metrics": { "downloads": 1645, "likes": 2, "lastModified": "2025-08-11" } }, { "id": "jais-2-70b-chat", "name": "Jais 2 70B Chat", "type": "llm", "country": "AE", "org": "Inception AI (G42)", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/inception42/Jais-2-70B-Chat", "paper": "https://arxiv.org/abs/2608.13580" }, "base_model": [ "from-scratch" ], "size": "70B", "year": 2025, "notes": "Second-generation Jais Arabic-English chat model released Dec 2025.", "metrics": { "downloads": 1628, "likes": 27, "lastModified": "2026-08-18" } }, { "id": "cohere-speech-tashkeel-2b", "name": "Cohere-Speech-Tashkeel-2B", "type": "asr", "country": "SA", "org": "NAMAA-Space", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr", "diacritization" ], "links": { "hf": "https://huggingface.co/NAMAA-Space/Cohere-Speech-Tashkeel-2B" }, "size": "2.1B", "year": 2026, "dialects": [ "msa" ], "notes": "Speech model that transcribes Arabic with tashkeel (diacritics).", "base_model": [ "coherelabs/cohere-transcribe-arabic-07-2026" ], "metrics": { "downloads": 1605, "likes": 28, "lastModified": "2026-07-13" } }, { "id": "deepseek-ocr-arabic-v1", "name": "deepseek ocr arabic v1", "type": "dataset", "country": "INTL", "org": "FatimahEmadEldin", "license": "unknown", "modality": "vision", "tasks": [ "ocr", "fine-tuning" ], "links": { "hf": "https://huggingface.co/datasets/FatimahEmadEldin/deepseek-ocr-arabic-v1" }, "size": "1K–10K rows", "year": 2026, "notes": "550 tokenized Arabic OCR samples with image patches and labels for fine-tuning DeepSeek-OCR.", "metrics": { "downloads": 1603, "likes": 0, "lastModified": "2026-02-09" } }, { "id": "rootformer-v14-sovereign-transmute", "name": "rootformer v14 sovereign transmute", "type": "llm", "country": "INTL", "org": "enver", "license": "other", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/enver/rootformer-v14-sovereign-transmute" }, "year": 2026, "tags": [ "variants:1" ], "notes": "Rootformer v14.3 formalizes four generative algorithmic (also: 1 variants)", "metrics": { "downloads": 1537, "likes": 0, "lastModified": "2026-09-27" } }, { "id": "arabic-game-arabizations", "name": "arabic game arabizations", "type": "dataset", "country": "INTL", "org": "jynxzio5", "license": "other", "modality": "text", "tasks": [ "nlp" ], "links": { "hf": "https://huggingface.co/datasets/jynxzio5/arabic-game-arabizations" }, "size": "10K–100K rows", "year": 2026, "notes": "Catalog scraped from wolfgame-ar.site for use by an installer app.", "metrics": { "downloads": 1535, "likes": 0, "lastModified": "2026-08-04" } }, { "id": "opus-mt-ar-fr", "name": "opus mt ar fr", "type": "llm", "country": "INTL", "org": "Helsinki-NLP", "license": "apache-2.0", "modality": "text", "tasks": [ "translation" ], "links": { "hf": "https://huggingface.co/Helsinki-NLP/opus-mt-ar-fr" }, "year": 2023, "notes": "OPUS-MT transformer translation model from Arabic to French.", "metrics": { "downloads": 1535, "likes": 1, "lastModified": "2023-08-16" } }, { "id": "oall-arabic-mmlu", "name": "OALL Arabic MMLU", "type": "benchmark", "country": "INTL", "org": "Open Arabic LLM Leaderboard (OALL)", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/OALL/Arabic_MMLU" }, "year": 2024, "dialects": [ "msa" ], "notes": "GPT-translated Arabic MMLU copy from FreedomIntelligence, used as an OALL v1 task.", "metrics": { "downloads": 1528, "likes": 3, "lastModified": "2024-09-05" } }, { "id": "jais-family-590m-chat", "name": "jais family 590m chat", "type": "llm", "country": "AE", "org": "Inception AI (G42)", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/inception42/jais-family-590m-chat" }, "base_model": [ "inceptionai/jais-family-590m" ], "size": "590M", "on_device": true, "year": 2024, "tags": [ "variants:3" ], "notes": "The Jais family of models is a comprehensive series of bilingual English-Arabic large language models (LLMs). (also: 3 variants)", "metrics": { "downloads": 1527, "likes": 7, "lastModified": "2024-09-11" } }, { "id": "nvidia-fastconformer-arabic-diacritics", "name": "NVIDIA FastConformer Arabic (Diacritics)", "type": "asr", "country": "INTL", "org": "NVIDIA", "license": "cc-by-4.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/nvidia/stt_ar_fastconformer_hybrid_large_pcd_v1.0" }, "notes": "Arabic ASR with diacritical marks, 1100h training", "metrics": { "downloads": 1524, "likes": 42, "lastModified": "2025-10-21" } }, { "id": "linto-dataset-audio-ar-tn", "name": "linto dataset audio ar tn", "type": "dataset", "country": "INTL", "org": "LINAGORA", "license": "cc-by-4.0", "modality": "speech", "tasks": [ "asr", "tts", "speech" ], "links": { "hf": "https://huggingface.co/datasets/linagora/linto-dataset-audio-ar-tn" }, "dialects": [ "magh" ], "size": "10K–100K rows", "year": 2025, "notes": "This is the first packaged version of the datasets used to train the Linto Tunisian dialect with code-switching STT (linagora/linto-asr-ar-tn).", "metrics": { "downloads": 1520, "likes": 22, "lastModified": "2025-04-11" } }, { "id": "human-translated-arabic-mmlu", "name": "Human-Translated Arabic MMLU", "type": "benchmark", "country": "AE", "org": "MBZUAI", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/MBZUAI/human_translated_arabic_mmlu" }, "year": 2024, "dialects": [ "msa" ], "notes": "Human-translated MMLU into Arabic from MBZUAI, companion to ArabicMMLU.", "metrics": { "downloads": 1514, "likes": 4, "lastModified": "2024-09-17" } }, { "id": "arabic-audio-collection-algerian-loubna-stories", "name": "arabic audio collection algerian loubna stories", "type": "dataset", "country": "EG", "org": "oddadmix", "license": "other", "modality": "speech", "tasks": [ "tts", "asr", "diacritization", "sentiment", "speech" ], "links": { "hf": "https://huggingface.co/datasets/oddadmix/arabic-audio-collection-algerian-loubna-stories" }, "dialects": [ "magh" ], "size": "10K–100K rows", "year": 2026, "notes": "The Loubna Stories Arabic Speech Dataset is a large-scale, first-of-its-kind Arabic speech corpus containing approximately 237 hours of speech recordings.", "metrics": { "downloads": 1478, "likes": 0, "lastModified": "2026-06-18" } }, { "id": "fanar-2-oryx-ivu", "name": "Fanar-2-Oryx-IVU", "type": "llm", "country": "QA", "org": "QCRI", "license": "apache-2.0", "modality": "multimodal", "tasks": [ "image-understanding", "chat" ], "links": { "hf": "https://huggingface.co/QCRI/Fanar-2-Oryx-IVU", "paper": "https://arxiv.org/abs/2603.16397" }, "size": "7B", "year": 2026, "notes": "Arabic-English vision-language model on Qwen2.5-VL-7B with cultural understanding focus.", "base_model": [ "qwen/qwen2.5-vl-7b-instruct" ], "metrics": { "downloads": 1463, "likes": 4, "lastModified": "2026-03-25" } }, { "id": "cc-100", "name": "CC-100", "type": "dataset", "country": "INTL", "org": "Facebook AI", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "website": "https://data.statmt.org/cc-100/", "hf": "https://huggingface.co/datasets/statmt/cc100", "paper": "https://aclanthology.org/2020.lrec-1.494.pdf" }, "dialects": [ "mixed" ], "size": "7,132,000 documents", "year": 2020, "notes": "Monolingual datasets from Common Crawl for a variety of languages", "metrics": { "downloads": 1462, "likes": 107, "lastModified": "2024-03-05" } }, { "id": "mt5-m2o-arabic-crosssum", "name": "mT5_m2o_arabic_crossSum", "type": "llm", "country": "INTL", "org": "csebuetnlp", "license": "unknown", "modality": "text", "tasks": [ "summarization" ], "links": { "hf": "https://huggingface.co/csebuetnlp/mT5_m2o_arabic_crossSum" }, "year": 2023, "notes": "This repository contains the many-to-one (m2o) mT5 checkpoint finetuned on all cross-lingual pairs of the CrossSum dataset.", "metrics": { "downloads": 1462, "likes": 4, "lastModified": "2023-11-15" } }, { "id": "wav2vec2-arabic-phoneme-asr", "name": "wav2vec2-arabic-phoneme-asr", "type": "asr", "country": "INTL", "org": "Mostafa Maroof", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/MostafaMaroof/wav2vec2-arabic-phoneme-asr" }, "size": "316M", "year": 2026, "on_device": true, "dialects": [ "msa" ], "notes": "wav2vec2 phoneme recognizer for Arabic.", "base_model": [ "facebook/wav2vec2-large-xlsr-53" ], "metrics": { "downloads": 1462, "likes": 1, "lastModified": "2026-07-06" } }, { "id": "tlog", "name": "tlog", "type": "dataset", "country": "INTL", "org": "Tarteel AI", "license": "cc-by-4.0", "modality": "speech", "tasks": [ "asr", "quran" ], "links": { "hf": "https://huggingface.co/datasets/tarteel-ai/tlog" }, "size": "100K–1M rows", "year": 2026, "notes": "TLOG: audio recitations of Quranic ayahs paired with their Quranic texts.", "metrics": { "downloads": 1441, "likes": 19, "lastModified": "2026-09-17" } }, { "id": "noormontai", "name": "NoormontAI", "type": "llm", "country": "INTL", "org": "Noormont", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/Noormont/NoormontAI" }, "year": 2026, "notes": "تم حفظ النموذج عند الخطوة 20000.", "base_model": [ "noormont/noormontai" ], "metrics": { "downloads": 1399, "likes": 0, "lastModified": "2026-09-09" } }, { "id": "whisper-tiny-ar-quran", "name": "whisper-tiny-ar-quran", "type": "asr", "country": "INTL", "org": "Tarteel AI", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/tarteel-ai/whisper-tiny-ar-quran" }, "year": 2022, "on_device": true, "dialects": [ "classical" ], "notes": "Whisper tiny fine-tuned for Quran recitation transcription.", "metrics": { "downloads": 1398, "likes": 29, "lastModified": "2022-12-14" } }, { "id": "arabart", "name": "AraBART", "type": "llm", "country": "INTL", "org": "Moussa Kamal Eddine", "license": "apache-2.0", "modality": "text", "tasks": [ "summarization", "pretraining" ], "links": { "hf": "https://huggingface.co/moussaKam/AraBART" }, "year": 2022, "dialects": [ "msa" ], "notes": "AraBART, first Arabic seq2seq model with pretrained encoder and decoder.", "metrics": { "downloads": 1391, "likes": 18, "lastModified": "2022-05-05" } }, { "id": "aramix", "name": "AraMix", "type": "dataset", "country": "INTL", "org": "AdaMLLab", "license": "other", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/datasets/AdaMLLab/AraMix" }, "size": "100M–1B rows", "year": 2026, "notes": "AraMix: Arabic pretraining corpus of over 100M rows combining deduplicated sources, with a quality-matched split; see arXiv 2512.18834.", "metrics": { "downloads": 1373, "likes": 7, "lastModified": "2026-01-30" } }, { "id": "whisper-large-v3-arabic-dialectal-v2", "name": "whisper-large-v3-arabic-dialectal-v2", "type": "asr", "country": "EG", "org": "oddadmix", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/oddadmix/whisper-large-v3-arabic-dialectal-v2" }, "size": "1.5B", "year": 2026, "dialects": [ "mixed" ], "notes": "Whisper large-v3 fine-tuned on Arabic dialects.", "base_model": [ "openai/whisper-large-v3" ], "metrics": { "downloads": 1370, "likes": 2, "lastModified": "2026-07-16" } }, { "id": "emiratitts-smoke-samples", "name": "EmiratiTTS-smoke-samples", "type": "dataset", "country": "INTL", "org": "Alqayed2024", "license": "mit", "modality": "speech", "tasks": [ "chat", "tts", "speech" ], "links": { "hf": "https://huggingface.co/datasets/Alqayed2024/EmiratiTTS-smoke-samples" }, "dialects": [ "gulf", "lev" ], "size": "<1K rows", "year": 2026, "notes": "These 10 audio clips are the stage 0.5 acceptance check for the EmiratiTTS project.", "metrics": { "downloads": 1366, "likes": 1, "lastModified": "2026-04-12" } }, { "id": "quranlab-arabic-speech", "name": "QuranLab-Arabic-Speech", "type": "dataset", "country": "INTL", "org": "Muno459", "license": "other", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/datasets/Muno459/QuranLab-Arabic-Speech" }, "size": "1M–10M rows", "year": 2026, "notes": "Broad Arabic speech from 50,000+ hours of recordings, cut into 2-20 s segments for ASR training; access restricted.", "metrics": { "downloads": 1347, "likes": 0, "lastModified": "2026-10-05" } }, { "id": "arabic-minilm-l12-v2-all-nli-triplet", "name": "Arabic-MiniLM-L12-v2-all-nli-triplet", "type": "embedding", "country": "SA", "org": "Omartificial-Intelligence-Space", "license": "apache-2.0", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/Omartificial-Intelligence-Space/Arabic-MiniLM-L12-v2-all-nli-triplet" }, "size": "118M", "year": 2024, "on_device": true, "dialects": [ "msa" ], "notes": "Small Arabic MiniLM sentence embedding model trained on NLI triplets.", "base_model": [ "sentence-transformers/paraphrase-multilingual-minilm-l12-v2" ], "metrics": { "downloads": 1338, "likes": 4, "lastModified": "2025-06-10" } }, { "id": "xtd-11", "name": "xtd 11", "type": "dataset", "country": "INTL", "org": "Arabic-Clip", "license": "unknown", "modality": "vision", "tasks": [ "ocr", "embedding", "vision" ], "links": { "hf": "https://huggingface.co/datasets/Arabic-Clip/xtd_11" }, "size": "1K–10K rows", "year": 2024, "notes": "The expanded XTD-11 dataset, now including Arabic, enhances the original XTD collection.", "metrics": { "downloads": 1326, "likes": 3, "lastModified": "2024-08-11" } }, { "id": "unnamed-ar", "name": "unnamed ar", "type": "dataset", "country": "EG", "org": "oddadmix", "license": "cc-by-3.0", "modality": "speech", "tasks": [ "asr", "speech" ], "links": { "hf": "https://huggingface.co/datasets/oddadmix/unnamed-ar" }, "size": "10K–100K rows", "year": 2026, "notes": "The data/ar/ partition of espnet/yodas3, repackaged as parquet with the audio inline so it loads without a script: Every row is one long-form recording.", "metrics": { "downloads": 1324, "likes": 0, "lastModified": "2026-09-29" } }, { "id": "karnak-40b-v1-0", "name": "Karnak 40B v1.0", "type": "llm", "country": "EG", "org": "Applied Innovation Center (MCIT Egypt)", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/Applied-Innovation-Center/Karnak-40B-v1.0" }, "size": "40B", "on_device": false, "year": 2026, "notes": "Karnak is a powerful AI model that works in both Arabic and English.", "base_model": [ "qwen/qwen3-30b-a3b-instruct-2507" ], "metrics": { "downloads": 1323, "likes": 72, "lastModified": "2026-04-24" } }, { "id": "karnak-llm", "name": "Karnak LLM", "type": "llm", "country": "EG", "org": "Applied Innovation Center (MCIT Egypt)", "license": "apache-2.0", "modality": "text", "tasks": [ "legal" ], "links": { "website": "https://itida.gov.eg/Arabic/PressReleases/Pages/egypt-national-ai-karnak-llm-launch-Ai-Everything-MEA-2026.aspx", "hf": "https://huggingface.co/Applied-Innovation-Center/Karnak" }, "year": 2026, "notes": "Egypt national Arabic LLM; first national apps built on it: SIA tutor and a legal/regulatory guidance assistant.", "base_model": [ "qwen/qwen3-30b-a3b-instruct-2507" ], "metrics": { "downloads": 1323, "likes": 72, "lastModified": "2026-04-24" } }, { "id": "silma-kashif", "name": "SILMA Kashif", "type": "llm", "country": "INTL", "org": "SILMA AI", "license": "gemma", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/silma-ai/SILMA-Kashif-2B-Instruct-v1.0" }, "size": "2B", "on_device": true, "dialects": [ "msa" ], "notes": "Lightweight RAG-optimized Arabic model", "metrics": { "downloads": 1315, "likes": 24, "lastModified": "2025-06-11" } }, { "id": "flair-arabic-multi-ner", "name": "flair arabic multi ner", "type": "llm", "country": "INTL", "org": "megantosh", "license": "apache-2.0", "modality": "text", "tasks": [ "encoder", "embedding", "ner" ], "links": { "hf": "https://huggingface.co/megantosh/flair-arabic-multi-ner" }, "year": 2022, "notes": "Training was conducted over 94 epochs, using a linear decaying learning rate of 2e-05, starting from 0.225 and a batch size of 32 with GloVe.", "metrics": { "downloads": 1297, "likes": 5, "lastModified": "2022-03-09" } }, { "id": "arabic-english-bge-m3", "name": "arabic english bge m3", "type": "embedding", "country": "INTL", "org": "sayed0am", "license": "mit", "modality": "text", "tasks": [ "embedding", "evaluation" ], "links": { "hf": "https://huggingface.co/sayed0am/arabic-english-bge-m3" }, "year": 2025, "notes": "This model is a 36.2% smaller version of BAAI/bge-m3 for the Arabic language.", "base_model": [ "baai/bge-m3" ], "metrics": { "downloads": 1282, "likes": 9, "lastModified": "2025-10-25" } }, { "id": "athar-shamela4", "name": "Athar Shamela4", "type": "dataset", "country": "INTL", "org": "Kandil7", "license": "mit", "modality": "text", "tasks": [ "ner", "qa" ], "links": { "hf": "https://huggingface.co/datasets/Kandil7/Athar-Shamela4" }, "size": "10M–100M rows", "year": 2026, "notes": "A complete extraction of al-Maktaba al-Shamela (الشاملة) v4, containing 8,589 books across 40 categories of classical Islamic sciences.", "metrics": { "downloads": 1262, "likes": 1, "lastModified": "2026-06-02" } }, { "id": "miracl-vision", "name": "MIRACL-VISION", "type": "benchmark", "country": "INTL", "org": "NVIDIA", "license": "cc-by-sa-4.0", "modality": "vision", "tasks": [ "evaluation", "retrieval", "rag" ], "links": { "hf": "https://huggingface.co/datasets/nvidia/miracl-vision", "paper": "https://arxiv.org/pdf/2505.11651" }, "dialects": [ "msa" ], "size": "75,444 documents", "year": 2025, "tags": [ "multilingual" ], "notes": "Multilingual visual document retrieval benchmark extending MIRACL with Wikipedia page images for 18 languages.", "metrics": { "downloads": 1255, "likes": 13, "lastModified": "2025-05-20" } }, { "id": "qari-ocr-0-2-2-1-vl-2b-instruct", "name": "Qari OCR 0.2.2.1 VL 2B Instruct", "type": "ocr", "country": "SA", "org": "NAMAA-Space", "license": "apache-2.0", "modality": "vision", "tasks": [ "chat", "ocr", "vision" ], "links": { "hf": "https://huggingface.co/NAMAA-Space/Qari-OCR-0.2.2.1-VL-2B-Instruct" }, "size": "2B", "on_device": true, "year": 2025, "notes": "This is the model described in the paper QARI-OCR: High-Fidelity Arabic Text Recognition through Multimodal Large Language Model Adaptation.", "base_model": [ "unsloth/qwen2-vl-2b-instruct-unsloth-bnb-4bit" ], "metrics": { "downloads": 1250, "likes": 25, "lastModified": "2025-06-07" } }, { "id": "moonshine-tiny-ar", "name": "moonshine tiny ar", "type": "asr", "country": "INTL", "org": "moonshine-ai", "license": "other", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/moonshine-ai/moonshine-tiny-ar" }, "year": 2025, "notes": "This model is part of the Moonshine family of tiny specialized Automatic Speech Recognition (ASR) models for edge devices.", "metrics": { "downloads": 1249, "likes": 8, "lastModified": "2025-09-06" } }, { "id": "bert-mini-arabic", "name": "bert mini arabic", "type": "llm", "country": "INTL", "org": "asafaya", "license": "unknown", "modality": "text", "tasks": [ "encoder" ], "links": { "hf": "https://huggingface.co/asafaya/bert-mini-arabic" }, "year": 2023, "tags": [ "variants:2" ], "notes": "Arabic BERT Mini encoder pretrained on OSCAR and Wikipedia.", "on_device": true, "metrics": { "downloads": 1244, "likes": 3, "lastModified": "2023-03-20" } }, { "id": "marbert-all-nli-triplet-matryoshka", "name": "Marbert all nli triplet Matryoshka", "type": "embedding", "country": "SA", "org": "Omartificial-Intelligence-Space", "license": "apache-2.0", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/Omartificial-Intelligence-Space/Marbert-all-nli-triplet-Matryoshka" }, "year": 2025, "notes": "This is a sentence-transformers model finetuned from UBC-NLP/MARBERTv2 on the Omartificial-Intelligence-Space/arabic-nli-triplet dataset.", "base_model": [ "ubc-nlp/marbertv2" ], "metrics": { "downloads": 1223, "likes": 1, "lastModified": "2025-01-10" } }, { "id": "ar-quran-hadith14books-msa", "name": "ar quran hadith14books MSA", "type": "dataset", "country": "INTL", "org": "Dr-AliGomaa", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr", "chat", "summarization", "speech", "quran", "hadith" ], "links": { "hf": "https://huggingface.co/datasets/Dr-AliGomaa/ar-quran-hadith14books-MSA" }, "dialects": [ "classical", "msa" ], "size": "10K–100K rows", "year": 2026, "notes": "Arabic speech for both primary sources of Islam — Quran and Hadith.", "metrics": { "downloads": 1219, "likes": 6, "lastModified": "2026-08-14" } }, { "id": "audar-asr-v1-flash", "name": "Audar ASR V1 Flash", "type": "asr", "country": "INTL", "org": "Audar AI", "license": "other", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/audarai/Audar-ASR-V1-Flash" }, "dialects": [ "gulf" ], "year": 2026, "notes": "Audar's Arabic-first ASR — the real-time, edge tier.", "metrics": { "downloads": 1216, "likes": 7, "lastModified": "2026-08-20" } }, { "id": "bert-base-arabic-camelbert-mix-did-madar-corpus6", "name": "bert base arabic camelbert mix did madar corpus6", "type": "llm", "country": "AE", "org": "CAMeL Lab, NYUAD", "license": "apache-2.0", "modality": "text", "tasks": [ "encoder", "pretraining" ], "links": { "hf": "https://huggingface.co/CAMeL-Lab/bert-base-arabic-camelbert-mix-did-madar-corpus6" }, "dialects": [ "mixed" ], "year": 2021, "notes": "For the fine-tuning, we used the MADAR Corpus 6 dataset, which includes 6 labels.", "metrics": { "downloads": 1216, "likes": 1, "lastModified": "2021-10-17" } }, { "id": "arbert", "name": "ARBERT", "type": "llm", "country": "INTL", "org": "UBC-NLP", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/UBC-NLP/ARBERT" }, "year": 2022, "dialects": [ "msa" ], "notes": "ARBERT, 61GB MSA BERT from UBC.", "metrics": { "downloads": 1208, "likes": 5, "lastModified": "2022-01-19" } }, { "id": "oall-arabic-exams", "name": "Arabic EXAMS (OALL)", "type": "benchmark", "country": "INTL", "org": "Open Arabic LLM Leaderboard (OALL)", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/OALL/Arabic_EXAMS" }, "year": 2024, "dialects": [ "msa" ], "notes": "Arabic subset of the EXAMS multilingual school-exam benchmark, an OALL task.", "metrics": { "downloads": 1207, "likes": 3, "lastModified": "2024-02-16" } }, { "id": "speecht5-tts-clartts-ar", "name": "speecht5_tts_clartts_ar", "type": "tts", "country": "AE", "org": "MBZUAI", "license": "cc-by-nc-4.0", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/MBZUAI/speecht5_tts_clartts_ar" }, "dialects": [ "classical" ], "notes": "SpeechT5 for Classical Arabic TTS", "metrics": { "downloads": 1198, "likes": 32, "lastModified": "2025-09-10" } }, { "id": "muffakir-embedding-v2", "name": "Muffakir Embedding V2", "type": "embedding", "country": "INTL", "org": "mohamed2811", "license": "unknown", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/mohamed2811/Muffakir_Embedding_V2" }, "year": 2025, "tags": [ "variants:1" ], "notes": "Muffakir This is the second version of the MuffakirEmbedding model. (also: 1 variants)", "base_model": [ "sayed0am/arabic-english-bge-m3" ], "metrics": { "downloads": 1194, "likes": 6, "lastModified": "2025-05-24" } }, { "id": "algerian-youtube-comments", "name": "Algerian Youtube Comments", "type": "dataset", "country": "INTL", "org": "Touati Kamel", "license": "mit", "modality": "text", "tasks": [ "pretraining", "classification" ], "links": { "hf": "https://huggingface.co/datasets/touati-kamel/Algerian-Youtube-Comments" }, "dialects": [ "magh" ], "size": "10K–100K rows", "year": 2026, "notes": "Algerian YouTube comments dataset with a comment-extraction pipeline, for Darija NLP and pretraining.", "metrics": { "downloads": 1193, "likes": 0, "lastModified": "2026-09-28" } }, { "id": "arabic-flicker-8k", "name": "Arabic Flicker 8k", "type": "dataset", "country": "INTL", "org": "Arabic-Clip", "license": "unknown", "modality": "vision", "tasks": [ "image-captioning", "multimodal" ], "links": { "hf": "https://huggingface.co/datasets/Arabic-Clip/Arabic_Flicker_8k" }, "size": "1K–10K rows", "year": 2024, "notes": "Arabic version of Flickr8k images for Arabic image-captioning and CLIP training.", "metrics": { "downloads": 1189, "likes": 0, "lastModified": "2024-07-09" } }, { "id": "arabic-dialects-gold20", "name": "arabic dialects gold20", "type": "dataset", "country": "INTL", "org": "TigreGotico", "license": "cc-by-4.0", "modality": "text", "tasks": [ "tts", "diacritization" ], "links": { "hf": "https://huggingface.co/datasets/TigreGotico/arabic-dialects-gold20" }, "dialects": [ "mixed" ], "size": "<1K rows", "year": 2026, "notes": "660 sentences across 33 Arabic lects, each with diacritized dialectal spelling, gold IPA, English gloss and phonetic feature tags.", "metrics": { "downloads": 1188, "likes": 0, "lastModified": "2026-07-20" } }, { "id": "qari-ocr-0-4-0-vl-4b-instruct", "name": "Qari-OCR-0.4.0-VL-4B-Instruct", "type": "ocr", "country": "SA", "org": "NAMAA-Space", "license": "apache-2.0", "modality": "vision", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/NAMAA-Space/Qari-OCR-0.4.0-VL-4B-Instruct" }, "size": "4B", "year": 2025, "dialects": [ "msa" ], "notes": "Qari OCR 0.4.0, 4B vision-language Arabic OCR model.", "base_model": [ "qwen/qwen3-vl-4b-instruct" ], "metrics": { "downloads": 1172, "likes": 10, "lastModified": "2026-02-01" } }, { "id": "coru", "name": "CORU", "type": "dataset", "country": "INTL", "org": "DataScienceUIBK", "license": "mit", "modality": "text", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/datasets/abdoelsayed/CORU" }, "size": "10K–100K rows", "year": 2025, "notes": "Multilingual OCR and information extraction from receipts remains challenging, particularly for complex scripts like Ar", "metrics": { "downloads": 1164, "likes": 15, "lastModified": "2025-06-19" } }, { "id": "arams-restore", "name": "AraMS-Restore", "type": "dataset", "country": "INTL", "org": "Archatext", "license": "unknown", "modality": "multimodal", "tasks": [ "asr", "ocr", "evaluation", "vision" ], "links": { "hf": "https://huggingface.co/datasets/Archatext/AraMS-Restore" }, "size": "<1K rows", "year": 2026, "notes": "177 line images cropped from real damaged pages of a historical Arabic manuscript (book09), each with its transcription.", "metrics": { "downloads": 1157, "likes": 0, "lastModified": "2026-08-04" } }, { "id": "mgb2-arabic", "name": "mgb2 arabic", "type": "dataset", "country": "INTL", "org": "MohamedRashad", "license": "unknown", "modality": "speech", "tasks": [ "asr", "speech" ], "links": { "hf": "https://huggingface.co/datasets/MohamedRashad/mgb2-arabic" }, "size": "100K–1M rows", "year": 2025, "notes": "The Arabic Multi-Genre Broadcast (MGB-2) dataset is a large-scale speech recognition corpus containing 1,200 hours of Arabic broadcast audio.", "metrics": { "downloads": 1156, "likes": 7, "lastModified": "2025-12-27" } }, { "id": "arabic-documents-dataset-pdf", "name": "Arabic Documents Dataset PDF", "type": "dataset", "country": "INTL", "org": "HumynLabs", "license": "cc-by-4.0", "modality": "vision", "tasks": [ "ocr", "document-understanding" ], "links": { "hf": "https://huggingface.co/datasets/HumynLabs/Arabic_Documents_Dataset_PDF" }, "size": "<1K rows", "year": 2025, "notes": "127 Arabic PDF documents (books, articles, reports) for document understanding and OCR research.", "metrics": { "downloads": 1136, "likes": 0, "lastModified": "2025-11-07" } }, { "id": "arabic-all-nli-triplet-matryoshka", "name": "Arabic-all-nli-triplet-Matryoshka", "type": "embedding", "country": "SA", "org": "Omartificial-Intelligence-Space", "license": "apache-2.0", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/Omartificial-Intelligence-Space/Arabic-all-nli-triplet-Matryoshka" }, "size": "278M", "year": 2024, "on_device": true, "dialects": [ "msa" ], "notes": "Arabic Matryoshka embeddings trained on translated all-nli triplets.", "base_model": [ "sentence-transformers/paraphrase-multilingual-mpnet-base-v2" ], "metrics": { "downloads": 1124, "likes": 6, "lastModified": "2025-01-23" } }, { "id": "lahgtna-v3-small", "name": "Lahgtna v3 Small", "type": "dataset", "country": "EG", "org": "oddadmix", "license": "unknown", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/datasets/oddadmix/lahgtna-v3-small" }, "size": "267h", "year": 2026, "dialects": [ "mixed" ], "notes": "Dialect-balanced Arabic ASR corpus, 54,600 clips over 13 dialects.", "metrics": { "downloads": 1107, "likes": 7, "lastModified": "2026-07-17" } }, { "id": "mms-300m-arabic-dialect-identifier", "name": "mms 300m arabic dialect identifier", "type": "asr", "country": "INTL", "org": "badrex", "license": "cc-by-4.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/badrex/mms-300m-arabic-dialect-identifier" }, "dialects": [ "mixed" ], "size": "300M", "on_device": true, "year": 2025, "notes": "Arabic audio classification model fine-tuned from facebook/mms-300m.", "base_model": [ "facebook/mms-300m" ], "metrics": { "downloads": 1080, "likes": 9, "lastModified": "2025-10-29" } }, { "id": "nabra-82m-v0-1", "name": "Nabra-82M-v0.1", "type": "tts", "country": "EG", "org": "oddadmix", "license": "apache-2.0", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/oddadmix/Nabra-82M-v0.1" }, "size": "82M", "year": 2026, "on_device": true, "dialects": [ "msa" ], "notes": "Nabra 82M compact Arabic text-to-speech.", "base_model": [ "hexgrad/kokoro-82m" ], "metrics": { "downloads": 1075, "likes": 10, "lastModified": "2026-07-03" } }, { "id": "nadi2024-baseline", "name": "NADI2024 baseline", "type": "llm", "country": "INTL", "org": "VARabi", "license": "unknown", "modality": "text", "tasks": [ "encoder", "embedding", "dialect-id" ], "links": { "hf": "https://huggingface.co/VARabi/NADI2024-baseline" }, "year": 2024, "notes": "A BERT-based model fine-tuned to perform single-label Arabic Dialect Identification (ADI).", "metrics": { "downloads": 1072, "likes": 2, "lastModified": "2024-09-22" } }, { "id": "stanza-ar", "name": "stanza ar", "type": "llm", "country": "INTL", "org": "stanfordnlp", "license": "apache-2.0", "modality": "text", "tasks": [ "encoder" ], "links": { "hf": "https://huggingface.co/stanfordnlp/stanza-ar" }, "year": 2026, "notes": "Stanza is a collection of accurate and efficient tools for the linguistic analysis of many human languages.", "metrics": { "downloads": 1066, "likes": 0, "lastModified": "2026-09-29" } }, { "id": "medqa-darija-multilingual", "name": "MedQA-Darija-MultiLingual", "type": "dataset", "country": "INTL", "org": "Williamsanderson", "license": "cc-by-4.0", "modality": "multimodal", "tasks": [ "medical", "qa", "asr" ], "links": { "hf": "https://huggingface.co/datasets/Williamsanderson/MedQA-Darija-MultiLingual" }, "dialects": [ "magh" ], "size": "100K–1M rows", "year": 2026, "notes": "Trilingual English, French and Moroccan Darija medical Q&A dataset with playable speech audio.", "metrics": { "downloads": 1063, "likes": 4, "lastModified": "2026-05-06" } }, { "id": "quran-md-ayahs", "name": "quran md ayahs", "type": "dataset", "country": "INTL", "org": "Buraaq", "license": "unknown", "modality": "speech", "tasks": [ "quran" ], "links": { "hf": "https://huggingface.co/datasets/Buraaq/quran-md-ayahs" }, "dialects": [ "classical" ], "size": "100K–1M rows", "year": 2026, "notes": "Ayah-level split of Quran-MD, a multimodal Quran dataset aligning text, linguistic annotation and recitation audio.", "metrics": { "downloads": 1054, "likes": 20, "lastModified": "2026-01-27" } }, { "id": "arabic-chat-llm-data", "name": "arabic chat llm data", "type": "dataset", "country": "INTL", "org": "amrykytt57", "license": "unknown", "modality": "text", "tasks": [ "chat", "diacritization" ], "links": { "hf": "https://huggingface.co/datasets/amrykytt57/arabic-chat-llm-data" }, "size": "10M–100M rows", "year": 2026, "notes": "Raw text from 4 sources (light cleaning: NFKC, diacritics/tatweel, URLs; exact de-duplication; no content classification), chat conversations, and token shards.", "metrics": { "downloads": 1048, "likes": 1, "lastModified": "2026-10-01" } }, { "id": "everyayah-wav", "name": "everyayah wav", "type": "dataset", "country": "INTL", "org": "Ahmed Hany", "license": "cc0-1.0", "modality": "speech", "tasks": [ "asr", "speech", "quran" ], "links": { "hf": "https://huggingface.co/datasets/dev-ahmedhany/everyayah-wav" }, "dialects": [ "classical" ], "size": "100K–1M rows", "year": 2026, "notes": "Full-mushaf Quranic recitation audio at 16 kHz mono 16-bit WAV, re-encoded from everyayah.com for ML / ASR research.", "metrics": { "downloads": 1043, "likes": 0, "lastModified": "2026-05-30" } }, { "id": "whisper-large-v3-turbo-arabic-dialectal-v2", "name": "whisper-large-v3-turbo-arabic-dialectal-v2", "type": "asr", "country": "EG", "org": "oddadmix", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/oddadmix/whisper-large-v3-turbo-arabic-dialectal-v2" }, "size": "809M", "year": 2026, "dialects": [ "mixed" ], "notes": "Whisper large-v3-turbo fine-tuned on Arabic dialects.", "base_model": [ "openai/whisper-large-v3-turbo" ], "metrics": { "downloads": 1038, "likes": 2, "lastModified": "2026-07-16" } }, { "id": "audar-tts-v1-flash", "name": "Audar TTS V1 Flash", "type": "tts", "country": "INTL", "org": "Audar AI", "license": "other", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/audarai/Audar-TTS-V1-Flash" }, "year": 2026, "notes": "Open, Arabic-first, expressive zero-shot text-to-speech — quantized to run anywhere.", "metrics": { "downloads": 1037, "likes": 9, "lastModified": "2026-07-07" } }, { "id": "gpt2-arabic-20k-lc", "name": "gpt2 arabic 20k lc", "type": "llm", "country": "INTL", "org": "aariciah", "license": "unknown", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/aariciah/gpt2-arabic-20k-lc" }, "year": 2026, "notes": "This model is a fine-tuned version of on the None dataset.", "metrics": { "downloads": 1012, "likes": 0, "lastModified": "2026-09-05" } }, { "id": "egymmlu", "name": "EgyMMLU", "type": "benchmark", "country": "INTL", "org": "UBC-NLP", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/UBC-NLP/EgyMMLU" }, "year": 2025, "dialects": [ "egy" ], "notes": "MMLU translated into Egyptian Arabic from UBC (Canada), released with NileChat.", "metrics": { "downloads": 996, "likes": 0, "lastModified": "2025-11-11" } }, { "id": "fastconformer-quran-ar", "name": "fastconformer quran ar", "type": "asr", "country": "INTL", "org": "mohammed", "license": "cc-by-4.0", "modality": "speech", "tasks": [ "asr", "quran" ], "links": { "hf": "https://huggingface.co/mohammed/fastconformer-quran-ar" }, "dialects": [ "classical" ], "year": 2026, "notes": "A fine-tuned NVIDIA FastConformer Hybrid Large model for Quranic Arabic speech recognition.", "metrics": { "downloads": 969, "likes": 18, "lastModified": "2026-06-07" } }, { "id": "petraai", "name": "PetraAI", "type": "dataset", "country": "INTL", "org": "PetraAI", "license": "apache-2.0", "modality": "text", "tasks": [ "ner", "qa", "translation", "summarization", "embedding", "tts" ], "links": { "hf": "https://huggingface.co/datasets/PetraAI/PetraAI" }, "size": "1M–10M rows", "year": 2023, "notes": "PETRA is a multilingual dataset for training and evaluating AI systems on a diverse range of tasks across multiple modalities.", "metrics": { "downloads": 947, "likes": 22, "lastModified": "2023-09-14" } }, { "id": "darijabert-arabizi", "name": "DarijaBERT-arabizi", "type": "llm", "country": "MA", "org": "SI2M Lab (INSEA)", "license": "unknown", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/SI2M-Lab/DarijaBERT-arabizi" }, "size": "255M", "dialects": [ "magh" ], "notes": "DarijaBERT variant for Darija written in Latin script (Arabizi).", "metrics": { "downloads": 943, "likes": 10, "lastModified": "2024-09-25" } }, { "id": "smolkalam-arabic-conversational-sft", "name": "smolkalam arabic conversational sft", "type": "dataset", "country": "INTL", "org": "AdaMLLab", "license": "apache-2.0", "modality": "text", "tasks": [ "chat", "translation", "instruction-tuning" ], "links": { "hf": "https://huggingface.co/datasets/AdaMLLab/smolkalam-arabic-conversational-sft" }, "size": "1M–10M rows", "year": 2026, "notes": "SmolKalam is a quality-filtered Arabic SFT dataset of 1,790,478 examples (2.45B tokens), built as an ensemble translation of SmolTalk2.", "metrics": { "downloads": 942, "likes": 3, "lastModified": "2026-08-17" } }, { "id": "nvidia-fastconformer-arabic", "name": "NVIDIA FastConformer Arabic", "type": "asr", "country": "INTL", "org": "NVIDIA", "license": "cc-by-4.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/nvidia/stt_ar_fastconformer_hybrid_large_pc_v1.0" }, "notes": "SOTA Arabic ASR, 115M params, 760h training, CC-BY-4.0", "metrics": { "downloads": 941, "likes": 19, "lastModified": "2025-10-23" } }, { "id": "arabic-whisper-codeswitching-edition", "name": "Arabic-Whisper-CodeSwitching-Edition", "type": "asr", "country": "INTL", "org": "MohamedRashad", "license": "gpl-3.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/MohamedRashad/Arabic-Whisper-CodeSwitching-Edition" }, "size": "1.5B", "year": 2024, "dialects": [ "mixed" ], "notes": "Whisper fine-tuned for Arabic with English code-switching.", "metrics": { "downloads": 939, "likes": 37, "lastModified": "2024-07-07" } }, { "id": "openiti-vectors", "name": "openiti vectors", "type": "dataset", "country": "INTL", "org": "Maktabati", "license": "cc-by-4.0", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/datasets/Maktabati/openiti-vectors" }, "size": "1M–10M rows", "year": 2026, "notes": "This dataset contains the fully vectorized OpenITI RELEASE 2025-1-9 collection of classical Islamic texts, prepared for semantic search (RAG).", "metrics": { "downloads": 937, "likes": 0, "lastModified": "2026-05-30" } }, { "id": "darijatts-v0-1-500m", "name": "DarijaTTS-v0.1-500M", "type": "tts", "country": "MA", "org": "Kandir Research", "license": "apache-2.0", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/KandirResearch/DarijaTTS-v0.1-500M" }, "size": "499M", "year": 2025, "on_device": true, "dialects": [ "magh" ], "notes": "DarijaTTS 500M Moroccan Darija text-to-speech.", "base_model": [ "outeai/outetts-0.2-500m" ], "metrics": { "downloads": 936, "likes": 0, "lastModified": "2025-11-26" } }, { "id": "moulsot-v0-3", "name": "moulsot.v0.3", "type": "asr", "country": "MA", "org": "atlasia", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/atlasia/moulsot.v0.3" }, "size": "2B", "year": 2026, "dialects": [ "magh" ], "notes": "Moulsot ASR for Moroccan Darija from Atlasia.", "base_model": [ "qwen/qwen3-asr-1.7b" ], "metrics": { "downloads": 936, "likes": 17, "lastModified": "2026-05-02" } }, { "id": "tarteel-ea-ud", "name": "Tarteel EA-UD", "type": "dataset", "country": "INTL", "org": "Tarteel AI", "license": "unknown", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/datasets/tarteel-ai/EA-UD" }, "year": 2022, "dialects": [ "classical" ], "notes": "Tarteel evaluation set of user-submitted Quran recitations.", "metrics": { "downloads": 936, "likes": 2, "lastModified": "2022-07-15" } }, { "id": "asr-code-switch", "name": "ASR Code Switch", "type": "benchmark", "country": "INTL", "org": "Perle-ai", "license": "mit", "modality": "speech", "tasks": [ "asr", "evaluation", "speech" ], "links": { "hf": "https://huggingface.co/datasets/Perle-ai/ASR_Code_Switch" }, "size": "1K–10K rows", "year": 2026, "notes": "A curated benchmark of 1,200 code-switching utterances (300 per language pair) for evaluating commercial ASR systems on multilingual speech.", "metrics": { "downloads": 923, "likes": 12, "lastModified": "2026-05-22" } }, { "id": "marbertv2-arabic-written-dialect-classifier", "name": "marbertv2 arabic written dialect classifier", "type": "llm", "country": "EG", "org": "Ibrahim Amin", "license": "apache-2.0", "modality": "text", "tasks": [ "encoder", "dialect-id" ], "links": { "hf": "https://huggingface.co/IbrahimAmin/marbertv2-arabic-written-dialect-classifier" }, "dialects": [ "egy", "magh", "msa", "mixed" ], "year": 2025, "notes": "This model is a fine-tuned version of UBC-NLP/MARBERTv2 for Arabic written dialect classification.", "base_model": [ "ubc-nlp/marbertv2" ], "metrics": { "downloads": 923, "likes": 6, "lastModified": "2025-05-11" } }, { "id": "namaa-saudi-tts-v2", "name": "NAMAA-Saudi-TTS-V2", "type": "tts", "country": "SA", "org": "NAMAA-Space", "license": "cc-by-nc-sa-4.0", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/NAMAA-Space/NAMAA-Saudi-TTS-V2" }, "year": 2026, "dialects": [ "gulf" ], "notes": "Second-generation Saudi-dialect TTS from NAMAA.", "base_model": [ "swivid/habibi-tts" ], "metrics": { "downloads": 922, "likes": 10, "lastModified": "2026-04-21" } }, { "id": "asl-4b-v1", "name": "ASL 4B v1", "type": "llm", "country": "INTL", "org": "saai-sa", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/saai-sa/ASL-4B-v1" }, "dialects": [ "gulf", "msa" ], "size": "4B", "on_device": false, "year": 2026, "notes": "ASL (أصل) is a 4-billion-parameter Arabic language model developed by SAAI.", "base_model": [ "qwen3.5" ], "metrics": { "downloads": 920, "likes": 19, "lastModified": "2026-08-26" } }, { "id": "opus-mt-fr-ar", "name": "opus mt fr ar", "type": "llm", "country": "INTL", "org": "Helsinki-NLP", "license": "apache-2.0", "modality": "text", "tasks": [ "translation" ], "links": { "hf": "https://huggingface.co/Helsinki-NLP/opus-mt-fr-ar" }, "year": 2023, "notes": "Tatoeba-Challenge transformer translating French into Arabic, with several Arabic varieties as targets.", "metrics": { "downloads": 919, "likes": 0, "lastModified": "2023-08-16" } }, { "id": "artelingo-dummy", "name": "artelingo dummy", "type": "benchmark", "country": "INTL", "org": "youssef101", "license": "mit", "modality": "multimodal", "tasks": [ "ocr", "sentiment", "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/youssef101/artelingo-dummy" }, "size": "10K–100K rows", "year": 2023, "notes": "ArtELingo is a benchmark and dataset introduced in a research paper aimed at promoting work on diversity across languages and cultures.", "metrics": { "downloads": 912, "likes": 2, "lastModified": "2023-07-23" } }, { "id": "quran-ayah-corpus", "name": "Quran Ayah Corpus", "type": "dataset", "country": "INTL", "org": "rabah2026", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr", "speech", "quran" ], "links": { "hf": "https://huggingface.co/datasets/rabah2026/Quran-Ayah-Corpus" }, "dialects": [ "classical" ], "size": "100K–1M rows", "year": 2025, "notes": "Ayah-Corpus is a large-scale, multi-reciter Arabic speech dataset meticulously curated for Automatic Speech Recognition (ASR) tasks.", "metrics": { "downloads": 912, "likes": 2, "lastModified": "2025-09-19" } }, { "id": "masrygpt-chat-1-5b", "name": "MasryGPT-Chat-1.5B", "type": "llm", "country": "INTL", "org": "ISLAM-PO", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/ISLAM-PO/MasryGPT-Chat-1.5B" }, "dialects": [ "egy" ], "size": "1.5B", "on_device": true, "year": 2026, "notes": "80,000 Egyptian terms • 2,500 steps • Loss 0.079 • 2.9GB 16-bit Merged English | [بالمصري](#-بالمصري---الوثائق-الاح", "base_model": [ "qwen/qwen2.5-1.5b-instruct" ], "metrics": { "downloads": 909, "likes": 1, "lastModified": "2026-09-25" } }, { "id": "arab-dialects-20-countries-3m", "name": "arab dialects 20 countries 3m", "type": "dataset", "country": "INTL", "org": "ISLAM-PO", "license": "cc-by-4.0", "modality": "text", "tasks": [ "nlp" ], "links": { "hf": "https://huggingface.co/datasets/ISLAM-PO/arab-dialects-20-countries-3m" }, "dialects": [ "egy", "lev", "mixed" ], "size": "1M–10M rows", "year": 2026, "notes": "The 3M target figure is a raw-repository claim and is not yet fully verified by the Hub index.", "metrics": { "downloads": 886, "likes": 0, "lastModified": "2026-09-25" } }, { "id": "aragpt2-medium", "name": "aragpt2-medium", "type": "llm", "country": "LB", "org": "AUB MIND Lab", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/aubmindlab/aragpt2-medium" }, "base_model": [ "from-scratch" ], "size": "394M", "year": 2022, "on_device": true, "dialects": [ "msa" ], "notes": "AraGPT2 medium, 370M Arabic GPT-2 from AUB.", "metrics": { "downloads": 882, "likes": 11, "lastModified": "2023-10-30" } }, { "id": "asjp-cerist", "name": "asjp cerist", "type": "dataset", "country": "INTL", "org": "DarjaCore", "license": "other", "modality": "text", "tasks": [ "retrieval", "ocr", "qa" ], "links": { "hf": "https://huggingface.co/datasets/DarjaCore/asjp-cerist" }, "size": "10K–100K rows", "year": 2026, "notes": "37,807 academic PDFs (579,683 pages) with 21,604 metadata records from Algeria's ASJP platform in Arabic, French and English.", "metrics": { "downloads": 869, "likes": 1, "lastModified": "2026-09-30" } }, { "id": "arabic-mmlu-10percent", "name": "Arabic MMLU 10percent", "type": "benchmark", "country": "INTL", "org": "arcee-globe", "license": "unknown", "modality": "text", "tasks": [ "mmlu", "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/arcee-globe/Arabic_MMLU-10percent" }, "size": "1K–10K rows", "year": 2024, "notes": "10 percent subset of Arabic MMLU multiple-choice questions across subjects.", "metrics": { "downloads": 850, "likes": 0, "lastModified": "2024-08-13" } }, { "id": "audar-tts-v1-turbo", "name": "Audar TTS V1 Turbo", "type": "tts", "country": "INTL", "org": "Audar AI", "license": "other", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/audarai/Audar-TTS-V1-Turbo" }, "year": 2026, "notes": "Open, Arabic-first, expressive zero-shot text-to-speech — the balanced production default.", "metrics": { "downloads": 850, "likes": 5, "lastModified": "2026-07-07" } }, { "id": "arabert-all-nli-triplet-matryoshka", "name": "Arabert all nli triplet Matryoshka", "type": "embedding", "country": "SA", "org": "Omartificial-Intelligence-Space", "license": "apache-2.0", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/Omartificial-Intelligence-Space/Arabert-all-nli-triplet-Matryoshka" }, "year": 2025, "notes": "This is a sentence-transformers model finetuned from aubmindlab/bert-base-arabertv02 on the Omartificial-Intelligence-Space/arabic-nli-triplet dataset.", "base_model": [ "aubmindlab/bert-base-arabertv02" ], "metrics": { "downloads": 846, "likes": 11, "lastModified": "2025-01-23" } }, { "id": "arabic-img2md", "name": "arabic-img2md", "type": "dataset", "country": "INTL", "org": "MohamedRashad", "license": "gpl-3.0", "modality": "vision", "tasks": [ "multimodal" ], "links": { "hf": "https://huggingface.co/datasets/MohamedRashad/arabic-img2md" }, "notes": "15K PDF pages paired with Markdown", "metrics": { "downloads": 840, "likes": 14, "lastModified": "2024-11-28" } }, { "id": "arabic-tts-spark", "name": "Arabic-TTS-Spark", "type": "tts", "country": "INTL", "org": "IbrahimSalah", "license": "fair-noncommercial-research-license", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/IbrahimSalah/Arabic-TTS-Spark" }, "dialects": [ "msa" ], "notes": "Spark TTS fine-tuned on 300h clean Arabic, MSA with diacritics", "base_model": [ "sparkaudio/spark-tts-0.5b" ], "metrics": { "downloads": 830, "likes": 30, "lastModified": "2025-11-27" } }, { "id": "arabic-xvector-embeddings", "name": "arabic xvector embeddings", "type": "dataset", "country": "INTL", "org": "herwoww", "license": "mit", "modality": "speech", "tasks": [ "tts", "embedding", "speech" ], "links": { "hf": "https://huggingface.co/datasets/herwoww/arabic_xvector_embeddings" }, "size": "<1K rows", "year": 2024, "notes": "There is one speaker embedding for each utterance in the validation set of both datasets.", "metrics": { "downloads": 820, "likes": 5, "lastModified": "2024-05-13" } }, { "id": "linto-dataset-audio-ar-tn-augmented", "name": "linto dataset audio ar tn augmented", "type": "dataset", "country": "INTL", "org": "LINAGORA", "license": "cc-by-4.0", "modality": "speech", "tasks": [ "asr", "tts", "speech" ], "links": { "hf": "https://huggingface.co/datasets/linagora/linto-dataset-audio-ar-tn-augmented" }, "dialects": [ "magh" ], "size": "100K–1M rows", "year": 2025, "notes": "This is the augmented datasets used to train the Linto Tunisian dialect with code-switching STT linagora/linto-asr-ar-tn.", "metrics": { "downloads": 817, "likes": 7, "lastModified": "2025-04-11" } }, { "id": "karsl-502-arabic-sign-language-v2", "name": "karsl 502 arabic sign language v2", "type": "dataset", "country": "INTL", "org": "FatimahEmadEldin", "license": "unknown", "modality": "vision", "tasks": [ "sign-language" ], "links": { "hf": "https://huggingface.co/datasets/FatimahEmadEldin/karsl-502-arabic-sign-language-v2" }, "year": 2026, "tags": [ "variants:1" ], "notes": "KArSL-502 Arabic sign language image frames organized by signer and sign folders.", "metrics": { "downloads": 816, "likes": 1, "lastModified": "2026-01-20" } }, { "id": "karnak-6b", "name": "Karnak-6B-v1.0", "type": "llm", "country": "EG", "org": "Applied Innovation Center (MCIT Egypt)", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/Applied-Innovation-Center/Karnak-6B-v1.0" }, "size": "5.9B", "year": 2026, "dialects": [ "msa", "egy" ], "notes": "6B small sibling of the Karnak national Egyptian LLM from the Applied Innovation Center, MCIT.", "base_model": [ "qwen/qwen3-4b-instruct-2507" ], "metrics": { "downloads": 815, "likes": 7, "lastModified": "2026-05-11" } }, { "id": "ayncoding-qwen3-8b-slim", "name": "ayncoding qwen3 8b slim", "type": "llm", "country": "INTL", "org": "enver", "license": "other", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/enver/ayncoding-qwen3-8b-slim" }, "size": "8B", "on_device": false, "year": 2026, "notes": "Unlike standard models trained on unstructured, noisy repositories, AynCoding-Qwen3-8B-Slim actively purges illogical pre-", "base_model": [ "qwen/qwen3-8b" ], "metrics": { "downloads": 807, "likes": 0, "lastModified": "2026-09-27" } }, { "id": "faster-whisper-base-ar-quran", "name": "faster whisper base ar quran", "type": "asr", "country": "INTL", "org": "OdyAsh", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr", "translation", "quran" ], "links": { "hf": "https://huggingface.co/OdyAsh/faster-whisper-base-ar-quran" }, "dialects": [ "classical" ], "year": 2025, "notes": "This model is a CTranslate2 version of tarteel-ai/whisper-base-ar-quran.", "base_model": [ "tarteel-ai/whisper-base-ar-quran", "openai/whisper-base" ], "metrics": { "downloads": 806, "likes": 6, "lastModified": "2025-06-24" } }, { "id": "mr-tydi", "name": "Mr. TyDi", "type": "dataset", "country": "INTL", "org": "David R. Cheriton School of Computer Science", "license": "apache-2.0", "modality": "text", "tasks": [ "retrieval" ], "links": { "github": "https://github.com/castorini/mr.tydi", "hf": "https://huggingface.co/datasets/castorini/mr-tydi", "paper": "https://arxiv.org/pdf/2108.08787.pdf" }, "dialects": [ "msa" ], "size": "16,573 sentences", "year": 2021, "tags": [ "multilingual" ], "notes": "Multilingual retrieval benchmark built from TyDi QA across eleven languages including Arabic.", "metrics": { "downloads": 806, "likes": 23, "lastModified": "2022-10-12" } }, { "id": "qwen2-5-7b-instruct-arabic-yt-merged", "name": "qwen2.5 7b instruct arabic yt merged", "type": "llm", "country": "INTL", "org": "HarithSami", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/HarithSami/qwen2.5-7b-instruct-arabic-yt-merged" }, "size": "7B", "on_device": false, "year": 2026, "tags": [ "variants:1" ], "notes": "This qwen2 model was trained 2x faster with Unsloth and Huggingface's TRL library. (also: 1 variants)", "base_model": [ "unsloth/qwen2.5-7b-instruct" ], "metrics": { "downloads": 796, "likes": 0, "lastModified": "2026-09-16" } }, { "id": "arabert-arabic-sentiment-analysis", "name": "AraBert-Arabic-Sentiment-Analysis", "type": "llm", "country": "INTL", "org": "PRAli22", "license": "apache-2.0", "modality": "text", "tasks": [ "encoder", "embedding", "sentiment", "evaluation" ], "links": { "hf": "https://huggingface.co/PRAli22/AraBert-Arabic-Sentiment-Analysis" }, "year": 2024, "notes": "Arabic text classification model.", "metrics": { "downloads": 789, "likes": 5, "lastModified": "2024-03-13" } }, { "id": "arvoice", "name": "ArVoice", "type": "dataset", "country": "AE", "org": "MBZUAI", "license": "cc-by-4.0", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/datasets/MBZUAI/ArVoice" }, "dialects": [ "msa" ], "notes": "Multi-speaker MSA corpus, 83h, 11 voices, diacritized", "metrics": { "downloads": 781, "likes": 33, "lastModified": "2025-10-31" } }, { "id": "arabictext-large", "name": "ArabicText-Large", "type": "dataset", "country": "INTL", "org": "Jr23xd23", "license": "apache-2.0", "modality": "text", "tasks": [ "nlp" ], "links": { "hf": "https://huggingface.co/datasets/Jr23xd23/ArabicText-Large" }, "notes": "743K articles for LLM training", "metrics": { "downloads": 773, "likes": 69, "lastModified": "2025-10-27" } }, { "id": "clartts", "name": "ClArTTS", "type": "dataset", "country": "AE", "org": "MBZUAI", "license": "cc-by-4.0", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/datasets/MBZUAI/ClArTTS" }, "dialects": [ "classical" ], "notes": "Classical Arabic TTS corpus by MBZUAI", "metrics": { "downloads": 769, "likes": 27, "lastModified": "2025-10-01" } }, { "id": "abjad-kids", "name": "Abjad Kids", "type": "dataset", "country": "INTL", "org": "Aziz-snoubra", "license": "mit", "modality": "speech", "tasks": [ "speech", "education" ], "links": { "hf": "https://huggingface.co/datasets/Aziz-snoubra/Abjad-Kids" }, "size": "10K–100K rows", "year": 2026, "notes": "Abjad-Kids: Arabic speech classification dataset of children's recordings of letters, numbers and colors for primary education.", "metrics": { "downloads": 761, "likes": 1, "lastModified": "2026-03-14" } }, { "id": "silma-embedding-matryoshka-v0-1", "name": "silma-embedding-matryoshka-v0.1", "type": "embedding", "country": "INTL", "org": "SILMA AI", "license": "apache-2.0", "modality": "text", "tasks": [ "embedding", "retrieval" ], "links": { "hf": "https://huggingface.co/silma-ai/silma-embedding-matryoshka-v0.1" }, "size": "135M", "year": 2024, "on_device": true, "dialects": [ "msa" ], "notes": "SILMA Arabic and English Matryoshka embedding model.", "base_model": [ "aubmindlab/bert-base-arabertv02" ], "metrics": { "downloads": 755, "likes": 14, "lastModified": "2025-05-15" } }, { "id": "arad", "name": "ArAD", "type": "benchmark", "country": "INTL", "org": "SpeechAntiSpoofingBenchmarks", "license": "odc-by", "modality": "speech", "tasks": [ "speech", "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/SpeechAntiSpoofingBenchmarks/ArAD" }, "dialects": [ "lev" ], "size": "1K–10K rows", "year": 2026, "notes": "Benchmark-ready packaging of the test split of the Arabic Audio Deepfake (ArAD) dataset: binary anti-spoofing on Arabic (primarily Levantine dialect) speech.", "metrics": { "downloads": 749, "likes": 1, "lastModified": "2026-06-27" } }, { "id": "bojji", "name": "bojji", "type": "embedding", "country": "SA", "org": "NAMAA-Space", "license": "mit", "modality": "text", "tasks": [ "static-embedding" ], "links": { "hf": "https://huggingface.co/NAMAA-Space/bojji" }, "year": 2025, "notes": "Lightweight, fast Arabic static embedding model built with Model2Vec.", "on_device": true, "metrics": { "downloads": 745, "likes": 2, "lastModified": "2025-06-16" } }, { "id": "camelbert-msa-sentiment", "name": "CAMeLBERT-MSA-Sentiment", "type": "llm", "country": "AE", "org": "CAMeL Lab, NYUAD", "license": "apache-2.0", "modality": "text", "tasks": [ "task-specific" ], "links": { "hf": "https://huggingface.co/CAMeL-Lab/bert-base-arabic-camelbert-msa-sentiment" }, "dialects": [ "msa" ], "notes": "Sentiment Analysis - Fine-tuned for MSA sentiment", "metrics": { "downloads": 742, "likes": 8, "lastModified": "2021-10-17" } }, { "id": "mbert2mbert-arabic-text-summarization", "name": "mbert2mbert arabic text summarization", "type": "llm", "country": "INTL", "org": "malmarjeh", "license": "unknown", "modality": "text", "tasks": [ "summarization" ], "links": { "hf": "https://huggingface.co/malmarjeh/mbert2mbert-arabic-text-summarization" }, "dialects": [ "msa" ], "year": 2023, "notes": "BERT2BERT abstractive summarizer initialized from mBERT, fine-tuned on 84,764 Arabic paragraph-summary pairs.", "metrics": { "downloads": 737, "likes": 12, "lastModified": "2023-07-01" } }, { "id": "openiti-chunked", "name": "openiti chunked", "type": "dataset", "country": "INTL", "org": "mittagessen", "license": "cc0-1.0", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/datasets/mittagessen/openiti_chunked" }, "size": "10M–100M rows", "year": 2025, "notes": "This dataset is derived from the 2023.1.8 release of the OpenITI corpus and is intended to pretrain small language models with short context lengths.", "metrics": { "downloads": 737, "likes": 1, "lastModified": "2025-06-16" } }, { "id": "mtrini-svl-1-1-merged", "name": "Mtrini SVL 1.1 Merged", "type": "llm", "country": "INTL", "org": "CompiwerAI", "license": "apache-2.0", "modality": "text", "tasks": [ "chat", "vision" ], "links": { "hf": "https://huggingface.co/CompiwerAI/Mtrini-SVL-1.1-Merged" }, "dialects": [ "magh" ], "year": 2026, "notes": "Mtrini-SVL-1.1 is an open-weight 8B model developed by Compiwer AI, based on Qwen3-VL-8B-Instruct.", "base_model": [ "qwen/qwen3-vl-8b-instruct" ], "metrics": { "downloads": 735, "likes": 0, "lastModified": "2026-09-25" } }, { "id": "ivis-400m-gpu", "name": "ivis 400m gpu", "type": "llm", "country": "INTL", "org": "nepetai", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/nepetai/ivis-400m-gpu" }, "size": "451M", "on_device": true, "year": 2026, "notes": "Hybrid LM of about 451M parameters mixing GQA Transformer, Mamba SSM and MoE layers, Arabic and English.", "metrics": { "downloads": 734, "likes": 0, "lastModified": "2026-08-23" } }, { "id": "whisper-medium-finetuned-sada-asr", "name": "whisper medium finetuned sada asr", "type": "asr", "country": "INTL", "org": "wageehkhad", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/wageehkhad/whisper-medium-finetuned-sada-asr" }, "dialects": [ "gulf" ], "year": 2026, "notes": "openai/whisper-medium fine-tuned on the full SADA22 dataset (420 hours of Saudi Arabic speech) for Arabic automatic speech recognition (ASR).", "base_model": [ "openai/whisper-medium" ], "metrics": { "downloads": 730, "likes": 3, "lastModified": "2026-03-05" } }, { "id": "adi20", "name": "ADI-20", "type": "dataset", "country": "INTL", "org": "Elyadata", "license": "unknown", "modality": "speech", "tasks": [ "dialect-id" ], "links": { "hf": "https://huggingface.co/datasets/ArabicSpeech/ADI20" }, "year": 2026, "dialects": [ "mixed" ], "notes": "Arabic dialect identification dataset and models extending ADI-17 to 20 countries (Tunisia/Elyadata).", "metrics": { "downloads": 727, "likes": 0, "lastModified": "2026-07-04" } }, { "id": "bert-base-arabic-camelbert-mix-did-madar-corpus26", "name": "bert-base-arabic-camelbert-mix-did-madar-corpus26", "type": "llm", "country": "AE", "org": "CAMeL Lab, NYUAD", "license": "apache-2.0", "modality": "text", "tasks": [ "dialect-id" ], "links": { "hf": "https://huggingface.co/CAMeL-Lab/bert-base-arabic-camelbert-mix-did-madar-corpus26" }, "year": 2022, "dialects": [ "mixed" ], "notes": "CAMeLBERT-Mix fine-tuned for 26-city dialect ID on MADAR.", "metrics": { "downloads": 723, "likes": 4, "lastModified": "2025-10-27" } }, { "id": "whisper-large-v3-turbo-arabic-dialectal", "name": "whisper large v3 turbo arabic dialectal", "type": "asr", "country": "EG", "org": "oddadmix", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr", "diacritization" ], "links": { "hf": "https://huggingface.co/oddadmix/whisper-large-v3-turbo-arabic-dialectal" }, "dialects": [ "mixed" ], "year": 2026, "notes": "Fine-tune of openai/whisper-large-v3-turbo (809M) for multi-dialect Arabic speech recognition (undiacritized output).", "base_model": [ "openai/whisper-large-v3-turbo" ], "metrics": { "downloads": 723, "likes": 4, "lastModified": "2026-07-09" } }, { "id": "arabict5-17gb-base", "name": "ArabicT5-17GB-base", "type": "llm", "country": "INTL", "org": "sultan", "license": "unknown", "modality": "text", "tasks": [ "text-to-text" ], "links": { "hf": "https://huggingface.co/sultan/ArabicT5-17GB-base" }, "year": 2023, "tags": [ "variants:4" ], "notes": "ArabicT5 pre-trained on Arabic Wikipedia, Marefa, Hindawi books and news (17GB).", "metrics": { "downloads": 716, "likes": 4, "lastModified": "2023-11-09" } }, { "id": "tunisian-proverbs-with-image-associations-a-cultural-and-linguistic-da", "name": "Tunisian Proverbs with Image Associations A Cultural and Linguistic Dataset", "type": "dataset", "country": "INTL", "org": "HabibaAbderrahim", "license": "cc-by-4.0", "modality": "multimodal", "tasks": [ "translation", "vision" ], "links": { "hf": "https://huggingface.co/datasets/HabibaAbderrahim/Tunisian-Proverbs-with-Image-Associations-A-Cultural-and-Linguistic-Dataset" }, "dialects": [ "magh" ], "size": "<1K rows", "year": 2025, "notes": "Tunisian proverbs with explanations, English translations and AI-generated image associations.", "metrics": { "downloads": 714, "likes": 0, "lastModified": "2025-09-02" } }, { "id": "arabart-qalb14-gec-ged-13", "name": "arabart-qalb14-gec-ged-13", "type": "llm", "country": "AE", "org": "CAMeL Lab, NYUAD", "license": "mit", "modality": "text", "tasks": [ "grammar-correction" ], "links": { "hf": "https://huggingface.co/CAMeL-Lab/arabart-qalb14-gec-ged-13" }, "year": 2023, "dialects": [ "msa" ], "notes": "AraBART fine-tuned for Arabic grammatical error correction on QALB-2014.", "metrics": { "downloads": 705, "likes": 3, "lastModified": "2024-01-09" } }, { "id": "ara-df-2026", "name": "ArA-DF-2026", "type": "dataset", "country": "INTL", "org": "Elyadata", "license": "other", "modality": "speech", "tasks": [ "speech" ], "links": { "hf": "https://huggingface.co/datasets/ArabicSpeech/ArA-DF-2026" }, "size": "100K–1M rows", "year": 2026, "notes": "ArA-DF-2026 is an Arabic speech deepfake detection dataset for binary audio classification.", "metrics": { "downloads": 694, "likes": 1, "lastModified": "2026-08-04" } }, { "id": "50m-2048-emhotob", "name": "50M 2048 Emhotob", "type": "llm", "country": "EG", "org": "oddadmix", "license": "apache-2.0", "modality": "text", "tasks": [ "chat", "pretraining" ], "links": { "hf": "https://huggingface.co/oddadmix/50M-2048-Emhotob" }, "size": "50M", "on_device": true, "year": 2026, "tags": [ "variants:10" ], "notes": "from scratch on 20 billion Arabic tokens with a 2048-token context window. (also: 10 variants)", "metrics": { "downloads": 689, "likes": 7, "lastModified": "2026-07-14" } }, { "id": "alexandria-dialect", "name": "Alexandria", "type": "dataset", "country": "INTL", "org": "UBC-NLP", "license": "cc-by-nc-4.0", "modality": "text", "tasks": [ "dialect-id", "instruction-tuning" ], "links": { "hf": "https://huggingface.co/datasets/UBC-NLP/alexandria" }, "size": "107k samples", "year": 2026, "dialects": [ "mixed" ], "notes": "Community-driven dialectal Arabic dataset covering 13 Arab countries and 11 domains.", "metrics": { "downloads": 688, "likes": 15, "lastModified": "2026-06-24" } }, { "id": "gpt2-small-arabic", "name": "gpt2 small arabic", "type": "llm", "country": "INTL", "org": "akhooli", "license": "unknown", "modality": "text", "tasks": [ "diacritization", "poetry" ], "links": { "hf": "https://huggingface.co/akhooli/gpt2-small-arabic" }, "year": 2023, "notes": "GPT2 model from Arabic Wikipedia dataset based on gpt2-small (using Fastai2).", "metrics": { "downloads": 685, "likes": 19, "lastModified": "2023-03-20" } }, { "id": "quran-tajweed-phonetics", "name": "quran tajweed phonetics", "type": "dataset", "country": "INTL", "org": "Quran Lab", "license": "other", "modality": "text", "tasks": [ "asr", "tts", "quran" ], "links": { "hf": "https://huggingface.co/datasets/Quran-Lab/quran-tajweed-phonetics" }, "dialects": [ "classical" ], "size": "10K–100K rows", "year": 2026, "notes": "The complete phonetic layer of the Quran in the riwaya of Hafs 'an 'Asim via tariq al-Shatibiyyah.", "metrics": { "downloads": 684, "likes": 4, "lastModified": "2026-10-01" } }, { "id": "fahadprimex-b27-v3", "name": "FahadPrimeX-b27-V3", "type": "llm", "country": "INTL", "org": "FahadPrimeX", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/FahadPrimeX/FahadPrimeX-b27-V3" }, "year": 2026, "tags": [ "variants:1" ], "notes": "Base model: Qwen/Qwen3.8-27B (declared with thanks). (also: 1 variants)", "metrics": { "downloads": 678, "likes": 0, "lastModified": "2026-09-21" } }, { "id": "arabic-speech-to-text", "name": "arabic speech to text", "type": "asr", "country": "INTL", "org": "maherghanem86", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/maherghanem86/arabic_speech_to_text" }, "year": 2025, "notes": "تم تطوير وتدريب هذا النموذج للتعرف التلقائي على الكلام للغة العربية.", "metrics": { "downloads": 677, "likes": 1, "lastModified": "2025-09-07" } }, { "id": "quranic-asr-benchmark", "name": "Quranic ASR Benchmark", "type": "benchmark", "country": "INTL", "org": "Quran Lab", "license": "other", "modality": "speech", "tasks": [ "asr", "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/Quran-Lab/quranic-asr-benchmark" }, "size": "under 1K clips", "year": 2026, "dialects": [ "classical" ], "notes": "Small benchmark set for evaluating ASR models on Quran recitation.", "metrics": { "downloads": 677, "likes": 5, "lastModified": "2026-08-26" } }, { "id": "mmcqa-semeval27", "name": "MMCQA-SemEval27", "type": "benchmark", "country": "QA", "org": "QCRI", "license": "cc-by-nc-4.0", "modality": "multimodal", "tasks": [ "qa", "evaluation", "speech", "vision" ], "links": { "hf": "https://huggingface.co/datasets/QCRI/MMCQA-SemEval27" }, "size": "10K–100K rows", "year": 2026, "notes": "This is the dataset for MMCultureQA, the SemEval 2027 shared task on culturally grounded visual question answering in English and Arabic.", "metrics": { "downloads": 675, "likes": 12, "lastModified": "2026-09-23" } }, { "id": "egyptian-arabic-asr-clean", "name": "Egyptian Arabic ASR Clean", "type": "dataset", "country": "INTL", "org": "MAdel121", "license": "unknown", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/datasets/MAdel121/arabic-egy-cleaned" }, "dialects": [ "egy" ], "notes": "~72 hours of Egyptian Arabic speech", "metrics": { "downloads": 668, "likes": 18, "lastModified": "2025-05-04" } }, { "id": "alrage", "name": "ALRAGE", "type": "benchmark", "country": "INTL", "org": "Open Arabic LLM Leaderboard (OALL)", "license": "unknown", "modality": "text", "tasks": [ "evaluation", "rag" ], "links": { "hf": "https://huggingface.co/datasets/OALL/ALRAGE" }, "year": 2025, "dialects": [ "msa" ], "notes": "Arabic retrieval-augmented generation evaluation set used in OALL v2.", "metrics": { "downloads": 667, "likes": 4, "lastModified": "2024-11-14" } }, { "id": "arabic-labse-matryoshka", "name": "Arabic-labse-Matryoshka", "type": "embedding", "country": "SA", "org": "Omartificial-Intelligence-Space", "license": "apache-2.0", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/Omartificial-Intelligence-Space/Arabic-labse-Matryoshka" }, "size": "471M", "year": 2024, "on_device": true, "dialects": [ "msa" ], "notes": "Arabic Matryoshka sentence embeddings based on LaBSE.", "base_model": [ "sentence-transformers/labse" ], "metrics": { "downloads": 667, "likes": 5, "lastModified": "2025-01-10" } }, { "id": "darija-asr-corpus", "name": "darija asr corpus", "type": "dataset", "country": "INTL", "org": "abnajlae", "license": "['mit', 'cc-by-4.0', 'cc-by-sa-4.0']", "modality": "speech", "tasks": [ "asr", "speech" ], "links": { "hf": "https://huggingface.co/datasets/abnajlae/darija-asr-corpus" }, "dialects": [ "magh" ], "size": "10K–100K rows", "year": 2026, "notes": "Arabizi (Latin-script) transcriptions of Moroccan Darija speech, produced for a Whisper fine-tuning pipeline (paper not yet published -- citation forthcoming).", "metrics": { "downloads": 663, "likes": 0, "lastModified": "2026-09-07" } }, { "id": "m2cqa", "name": "M2CQA", "type": "dataset", "country": "QA", "org": "QCRI", "license": "cc-by-nc-sa-4.0", "modality": "multimodal", "tasks": [ "vision", "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/QCRI/M2CQA" }, "dialects": [ "egy", "lev", "msa" ], "size": "10K–100K rows", "year": 2026, "notes": "This release contains English, Modern Standard Arabic, Levantine Arabic, and Egyptian Arabic statements over a shared set of images.", "metrics": { "downloads": 662, "likes": 2, "lastModified": "2026-06-12" } }, { "id": "vynis-0-1-2b-instant", "name": "Vynis 0.1 2b instant", "type": "llm", "country": "INTL", "org": "OpenWeightsAI", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/OpenWeightsAI/Vynis-0.1-2b-instant" }, "on_device": true, "year": 2026, "notes": "Arabic-English instruction-tuned chat model fine-tuned from Llama-3.2-1B-Instruct on alpaca-cleaned.", "base_model": [ "unsloth/llama-3.2-1b-instruct" ], "metrics": { "downloads": 661, "likes": 0, "lastModified": "2026-09-15" } }, { "id": "aratrust", "name": "AraTrust", "type": "benchmark", "country": "INTL", "org": "Asas AI", "license": "mit", "modality": "text", "tasks": [ "evaluation", "safety" ], "links": { "hf": "https://huggingface.co/datasets/asas-ai/AraTrust" }, "size": "522 questions", "year": 2024, "dialects": [ "msa" ], "notes": "Arabic LLM trustworthiness benchmark across truthfulness, ethics, safety and privacy.", "metrics": { "downloads": 653, "likes": 4, "lastModified": "2024-05-07" } }, { "id": "darijabert", "name": "DarijaBERT", "type": "llm", "country": "MA", "org": "SI2M Lab (INSEA)", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/SI2M-Lab/DarijaBERT" }, "size": "209M", "year": 2022, "on_device": true, "dialects": [ "magh" ], "notes": "DarijaBERT, first BERT model for Moroccan Darija.", "metrics": { "downloads": 643, "likes": 37, "lastModified": "2024-09-25" } }, { "id": "arabic-openhermes-2-5", "name": "Arabic-OpenHermes-2.5", "type": "dataset", "country": "INTL", "org": "2A2I", "license": "apache-2.0", "modality": "text", "tasks": [ "nlp" ], "links": { "hf": "https://huggingface.co/datasets/2A2I/Arabic-OpenHermes-2.5" }, "notes": "Arabic OpenHermes instruction dataset", "metrics": { "downloads": 641, "likes": 21, "lastModified": "2024-03-15" } }, { "id": "arabic-sentiment-twitter-corpus-arbml", "name": "Arabic Sentiment Twitter Corpus", "type": "dataset", "country": "SA", "org": "ARBML", "license": "unknown", "modality": "text", "tasks": [ "sentiment" ], "links": { "hf": "https://huggingface.co/datasets/arbml/Arabic_Sentiment_Twitter_Corpus" }, "size": "10K–100K rows", "year": 2024, "notes": "58.7K Arabic tweets labelled for sentiment.", "metrics": { "downloads": 637, "likes": 2, "lastModified": "2024-03-30" } }, { "id": "arabic-oscar", "name": "Arabic OSCAR", "type": "dataset", "country": "INTL", "org": "Inria", "license": "cc0-1.0", "modality": "text", "tasks": [ "pretraining" ], "links": { "website": "https://oscar-corpus.com/", "hf": "https://huggingface.co/datasets/oscar-corpus/oscar", "paper": "https://arxiv.org/pdf/2006.06202.pdf" }, "dialects": [ "mixed" ], "size": "8,117,162,828 tokens", "year": 2020, "notes": "A huge multilingual corpus obtained by language classification and filtering of the Common Crawl", "metrics": { "downloads": 632, "likes": 208, "lastModified": "2025-09-04" } }, { "id": "nile-chat-4b", "name": "Nile-Chat-4B", "type": "llm", "country": "AE", "org": "MBZUAI-Paris Lab", "license": "gemma", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/MBZUAI-Paris/Nile-Chat-4B" }, "base_model": [ "google/gemma-3-4b-pt" ], "size": "3.9B", "year": 2025, "dialects": [ "egy" ], "notes": "Nile-Chat 4B, Egyptian dialect model handling both Arabic and Arabizi script.", "metrics": { "downloads": 617, "likes": 20, "lastModified": "2025-07-10" } }, { "id": "fikr-7b-reasoning", "name": "Fikr 7B Reasoning", "type": "llm", "country": "INTL", "org": "Hatim2221", "license": "apache-2.0", "modality": "text", "tasks": [ "reasoning", "chat" ], "links": { "hf": "https://huggingface.co/Hatim2221/Fikr-7B-Reasoning" }, "size": "7B", "on_device": false, "year": 2026, "notes": "Qwen2.5-7B-Instruct fine-tuned for structured chain-of-thought Arabic reasoning, evaluated on Arabic-GSM8K.", "base_model": [ "qwen/qwen2.5-7b-instruct" ], "metrics": { "downloads": 600, "likes": 1, "lastModified": "2026-08-29" } }, { "id": "livekit-turn-detector-arabic", "name": "livekit turn detector arabic", "type": "llm", "country": "INTL", "org": "Moustafa3092", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/Moustafa3092/livekit-turn-detector-arabic" }, "year": 2025, "notes": "Fine-tuned Arabic End-of-Utterance (EOU) detection model for LiveKit voice agents.", "base_model": [ "livekit/turn-detector" ], "metrics": { "downloads": 597, "likes": 3, "lastModified": "2025-12-14" } }, { "id": "aradice-arabicmmlu-egy", "name": "AraDICE ArabicMMLU Egyptian", "type": "benchmark", "country": "QA", "org": "QCRI", "license": "cc-by-nc-sa-4.0", "modality": "text", "tasks": [ "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/QCRI/AraDICE-ArabicMMLU-egy" }, "year": 2024, "dialects": [ "egy" ], "notes": "Egyptian-dialect translation of ArabicMMLU from the QCRI AraDiCE benchmark suite.", "metrics": { "downloads": 591, "likes": 1, "lastModified": "2024-11-08" } }, { "id": "jev-ar-bench", "name": "jev ar bench", "type": "benchmark", "country": "INTL", "org": "atmaneayoub", "license": "cc-by-4.0", "modality": "text", "tasks": [ "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/atmaneayoub/jev-ar-bench" }, "dialects": [ "gulf" ], "size": "1K–10K rows", "year": 2026, "notes": "Arabic intent-routing benchmarks for MSA, Emirati, Saudi and code-switched Arabic", "metrics": { "downloads": 591, "likes": 1, "lastModified": "2026-10-01" } }, { "id": "arabic-dialects-gold20-code-switch", "name": "arabic dialects gold20 code switch", "type": "dataset", "country": "INTL", "org": "TigreGotico", "license": "cc-by-4.0", "modality": "text", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/datasets/TigreGotico/arabic-dialects-gold20-code-switch" }, "dialects": [ "mixed" ], "size": "<1K rows", "year": 2026, "notes": "Code-switched Arabic sentences with IPA: 20 rows per lect across 33 Arabic lects (the same roster as the sibling TigreGotico/arabic-dialects-gold20).", "metrics": { "downloads": 582, "likes": 0, "lastModified": "2026-07-20" } }, { "id": "spark-tts-normazlied-masri-mega", "name": "spark tts normazlied masri mega", "type": "llm", "country": "EG", "org": "Mohamed Gomaa", "license": "unknown", "modality": "text", "tasks": [ "chat", "tts", "instruction-tuning" ], "links": { "hf": "https://huggingface.co/MohamedGomaa30/spark-tts-normazlied-masri-mega" }, "dialects": [ "egy" ], "year": 2026, "notes": "This model is a fine-tuned version of MohamedGomaa30/spark-tts-normazlied-masri-mega.", "base_model": [ "mohamedgomaa30/spark-tts-normazlied-masri-mega" ], "metrics": { "downloads": 581, "likes": 0, "lastModified": "2026-02-03" } }, { "id": "arabic-ner-pii2", "name": "Arabic NER PII2", "type": "llm", "country": "INTL", "org": "MutazYoune", "license": "apache-2.0", "modality": "text", "tasks": [ "encoder", "ner" ], "links": { "hf": "https://huggingface.co/MutazYoune/Arabic-NER-PII2" }, "year": 2025, "notes": "This is an Arabic Named Entity Recognition (NER) model fine-tuned on BERT architecture specifically for Arabic text processing.", "metrics": { "downloads": 579, "likes": 1, "lastModified": "2025-06-12" } }, { "id": "iraqi-sales-dialogue", "name": "Iraqi Arabic Sales Dialogue", "type": "dataset", "country": "IQ", "org": "Ameer Wisam", "license": "cc-by-4.0", "modality": "text", "tasks": [ "instruction-tuning" ], "links": { "hf": "https://huggingface.co/datasets/ameer4wisam/iraqi-arabic-sales-dialogue-dataset" }, "dialects": [ "iraqi" ], "notes": "Iraqi Arabic sales conversation dataset (Iraq).", "metrics": { "downloads": 577, "likes": 0, "lastModified": "2026-07-30" } }, { "id": "whisper-large-v3-egyptian-arabic", "name": "whisper-large-v3-egyptian-arabic", "type": "asr", "country": "EG", "org": "Abdelrahman Hassan", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/AbdelrahmanHassan/whisper-large-v3-egyptian-arabic" }, "year": 2025, "dialects": [ "egy" ], "notes": "Whisper large-v3 fine-tuned on Egyptian Arabic.", "base_model": [ "openai/whisper-large-v3" ], "metrics": { "downloads": 575, "likes": 12, "lastModified": "2025-08-04" } }, { "id": "atlas-chat-9b", "name": "Atlas-Chat-9B", "type": "llm", "country": "AE", "org": "MBZUAI-Paris Lab", "license": "gemma", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/MBZUAI-Paris/Atlas-Chat-9B" }, "base_model": [ "google/gemma-2-9b-it" ], "size": "9.2B", "year": 2024, "dialects": [ "magh" ], "notes": "Atlas-Chat 9B, Gemma-2 based instruction model for Moroccan Darija.", "metrics": { "downloads": 571, "likes": 30, "lastModified": "2024-10-24" } }, { "id": "covost-2", "name": "CoVoST 2", "type": "dataset", "country": "INTL", "org": "Facebook AI", "license": "cc0-1.0", "modality": "speech", "tasks": [ "asr", "translation" ], "links": { "github": "https://github.com/facebookresearch/covost", "hf": "https://huggingface.co/datasets/facebook/covost2", "paper": "https://arxiv.org/pdf/2007.10310.pdf" }, "dialects": [ "msa" ], "size": "6 hours", "year": 2020, "tags": [ "multilingual" ], "notes": "Multilingual speech translation corpus based on Common Voice, with Arabic among its languages.", "metrics": { "downloads": 567, "likes": 51, "lastModified": "2024-01-18" } }, { "id": "fa-en-ar-handwritten-ocr-v1", "name": "fa en ar handwritten ocr v1", "type": "benchmark", "country": "INTL", "org": "saeid1999", "license": "cc-by-4.0", "modality": "multimodal", "tasks": [ "ocr", "asr", "evaluation", "vision" ], "links": { "hf": "https://huggingface.co/datasets/saeid1999/fa-en-ar-handwritten-ocr-v1" }, "size": "10K–100K rows", "year": 2026, "notes": "A large, clean, augmentation-rich synthetic handwriting dataset for training and benchmarking OCR / HTR models on Persian (fa), Arabic (ar) and English (en).", "metrics": { "downloads": 565, "likes": 0, "lastModified": "2026-09-17" } }, { "id": "badr-embedding-v0", "name": "Badr embedding v0", "type": "embedding", "country": "INTL", "org": "somayaeltanbouly", "license": "unknown", "modality": "text", "tasks": [ "retrieval", "fiqh" ], "links": { "hf": "https://huggingface.co/somayaeltanbouly/Badr_embedding_v0" }, "year": 2026, "notes": "Arabic dense-retrieval embedding model for fiqh (Islamic jurisprudence), fine-tuned from Muffakir_Embedding.", "base_model": [ "mohamed2811/muffakir_embedding" ], "metrics": { "downloads": 556, "likes": 0, "lastModified": "2026-09-14" } }, { "id": "qween7-5-arabic-story-teller-2", "name": "qween7.5 arabic story teller 2", "type": "llm", "country": "INTL", "org": "Raido", "license": "apache-2.0", "modality": "text", "tasks": [ "story-generation" ], "links": { "hf": "https://huggingface.co/Raido/qween7.5-arabic-story-teller-2" }, "year": 2025, "notes": "Qwen2.5-7B-Instruct fine-tuned with Unsloth as an Arabic story-telling model.", "size": "7B", "base_model": [ "unsloth/qwen2.5-7b-instruct-bnb-4bit" ], "metrics": { "downloads": 556, "likes": 0, "lastModified": "2025-01-07" } }, { "id": "arabicbert-arabic-dialect-identification", "name": "arabicBert_arabic_dialect_identification", "type": "llm", "country": "INTL", "org": "lafifi-24", "license": "unknown", "modality": "text", "tasks": [ "encoder", "embedding", "dialect-id", "speech" ], "links": { "hf": "https://huggingface.co/lafifi-24/arabicBert_arabic_dialect_identification" }, "dialects": [ "mixed" ], "year": 2023, "notes": "Our Arabic Dialect Identification models are trained to accurately identify spoken dialects in Arabic text.", "metrics": { "downloads": 555, "likes": 0, "lastModified": "2023-04-30" } }, { "id": "voho-saudi-chat-4b", "name": "voho-saudi-chat-4b", "type": "llm", "country": "SA", "org": "Voho AI", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/VohoAI/voho-saudi-chat-4b" }, "size": "4B", "year": 2026, "dialects": [ "gulf" ], "notes": "Voho Saudi chat 4B, Saudi-dialect conversational model.", "base_model": [ "qwen/qwen3-4b-instruct-2507" ], "metrics": { "downloads": 555, "likes": 1, "lastModified": "2026-09-19" } }, { "id": "lahgtna-omnivoice-v2", "name": "Lahgtna OmniVoice v2", "type": "tts", "country": "EG", "org": "oddadmix", "license": "unknown", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/oddadmix/lahgtna-omnivoice-v2" }, "size": "613M", "year": 2026, "dialects": [ "mixed" ], "notes": "Multi-dialect Arabic TTS from the Lahgtna project, trained on dialectal speech.", "base_model": [ "k2-fsa/omnivoice" ], "metrics": { "downloads": 554, "likes": 13, "lastModified": "2026-06-15" } }, { "id": "alm-bench", "name": "ALM-Bench", "type": "benchmark", "country": "AE", "org": "MBZUAI", "license": "cc-by-nc-4.0", "modality": "multimodal", "tasks": [ "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/MBZUAI/ALM-Bench" }, "year": 2024, "notes": "All Languages Matter multimodal cultural benchmark covering 100 languages including Arabic.", "metrics": { "downloads": 553, "likes": 20, "lastModified": "2025-02-28" } }, { "id": "sfc-mini", "name": "SFC-mini", "type": "dataset", "country": "SA", "org": "King Saud University", "license": "cc-by-nc-4.0", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/datasets/faisalq/SFC-mini" }, "year": 2024, "dialects": [ "gulf" ], "notes": "Saudi forums corpus subset used to pretrain SaudiBERT.", "metrics": { "downloads": 551, "likes": 2, "lastModified": "2024-05-08" } }, { "id": "qwen3-4b-algerian-darja", "name": "Qwen3 4B Algerian Darja", "type": "llm", "country": "INTL", "org": "DarjaCore", "license": "['apache-2.0']", "modality": "text", "tasks": [ "chat", "pretraining" ], "links": { "hf": "https://huggingface.co/DarjaCore/Qwen3-4B-Algerian-Darja" }, "base_model": [ "qwen/qwen3-4b-base" ], "dialects": [ "magh" ], "size": "4B", "on_device": false, "year": 2026, "notes": "This model is an experimental Algerian Darja adaptation of Qwen/Qwen3-4B-Base.", "metrics": { "downloads": 548, "likes": 1, "lastModified": "2026-09-10" } }, { "id": "dclm-pro-arabic", "name": "dclm pro arabic", "type": "dataset", "country": "INTL", "org": "SultanR", "license": "unknown", "modality": "text", "tasks": [ "translation", "pretraining" ], "links": { "hf": "https://huggingface.co/datasets/SultanR/dclm-pro-arabic" }, "size": "10M–100M rows", "year": 2026, "notes": "Arabic translation of DCLM-Pro (global shards 01 and 05), translated with Seed-X-PPO-7B using greedy decoding.", "metrics": { "downloads": 547, "likes": 0, "lastModified": "2026-08-17" } }, { "id": "madlad-arabic-clean", "name": "MADLAD Arabic Clean", "type": "dataset", "country": "INTL", "org": "ymoslem", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/datasets/ymoslem/MADLAD-Arabic-Clean" }, "size": "10M–100M rows", "year": 2024, "notes": "Cleaned Arabic subset of MADLAD-400 with 12.4M documents (about 77 GB of text).", "metrics": { "downloads": 547, "likes": 0, "lastModified": "2024-07-11" } }, { "id": "jais-family-chat", "name": "Jais family chat models", "type": "llm", "country": "AE", "org": "Inception AI (G42)", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/inception42/jais-family-13b-chat" }, "base_model": [ "inceptionai/jais-family-13b" ], "size": "590M-30B", "year": 2024, "notes": "Jais family of bilingual Arabic-English chat models from 590M to 30B trained from scratch.", "metrics": { "downloads": 541, "likes": 10, "lastModified": "2024-09-11" } }, { "id": "evarest", "name": "EvArEST", "type": "dataset", "country": "INTL", "org": "Melaraby", "license": "bsd-3-clause", "modality": "vision", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/datasets/Melaraby/EvArEST-dataset-for-Arabic-scene-text-recognition" }, "year": 2022, "dialects": [ "msa" ], "notes": "Everyday Arabic-English scene text recognition dataset.", "metrics": { "downloads": 539, "likes": 2, "lastModified": "2025-11-09" } }, { "id": "ayncoding-qwen2-5-coder-1-5b", "name": "ayncoding qwen2.5 coder 1.5b", "type": "llm", "country": "INTL", "org": "enver", "license": "other", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/enver/ayncoding-qwen2.5-coder-1.5b" }, "size": "1.5B", "on_device": true, "year": 2026, "notes": "It serves as the ultra-fast Drafter model within the Hierarchical Symbolic-Neural MoE (H-MoE) architecture, generating high-velocity syntax.", "base_model": [ "qwen/qwen2.5-coder-1.5b" ], "metrics": { "downloads": 535, "likes": 0, "lastModified": "2026-09-27" } }, { "id": "synthetic-bilingual-invoices-200", "name": "synthetic bilingual invoices 200", "type": "dataset", "country": "INTL", "org": "HV09", "license": "cc-by-4.0", "modality": "multimodal", "tasks": [ "ocr", "vision" ], "links": { "hf": "https://huggingface.co/datasets/HV09/synthetic-bilingual-invoices-200" }, "size": "<1K rows", "year": 2026, "notes": "200 rendered invoice images and a matching 17-field ground-truth record for every one.", "metrics": { "downloads": 535, "likes": 0, "lastModified": "2026-08-01" } }, { "id": "idiomx", "name": "IdiomX", "type": "benchmark", "country": "INTL", "org": "aymansharara", "license": "mit", "modality": "text", "tasks": [ "embedding", "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/aymansharara/IdiomX" }, "size": "100K–1M rows", "year": 2026, "notes": "Dataset: 190K+ examples • 12K+ idioms • 3 languages • 4 benchmark tasks It supports four benchmark tasks: 1.", "metrics": { "downloads": 534, "likes": 2, "lastModified": "2026-06-05" } }, { "id": "muharaf-public", "name": "muharaf public", "type": "dataset", "country": "INTL", "org": "Muharaf Project", "license": "cc-by-nc-sa-4.0", "modality": "multimodal", "tasks": [ "ocr", "vision" ], "links": { "hf": "https://huggingface.co/datasets/aamijar/muharaf-public" }, "size": "10K–100K rows", "year": 2025, "notes": "This dataset contains 24,495 line images of Arabic handwriting and the corresponding text.", "metrics": { "downloads": 534, "likes": 21, "lastModified": "2025-01-24" } }, { "id": "arabic-quran-asr-dataset", "name": "Arabic Quran ASR dataset", "type": "dataset", "country": "INTL", "org": "Sabri12blm", "license": "unknown", "modality": "speech", "tasks": [ "asr", "quran" ], "links": { "hf": "https://huggingface.co/datasets/Sabri12blm/Arabic-Quran-ASR-dataset" }, "dialects": [ "classical" ], "size": "10K–100K rows", "year": 2025, "notes": "Quranic recitation ASR dataset of 74,832 audio-transcript pairs.", "metrics": { "downloads": 531, "likes": 5, "lastModified": "2025-02-19" } }, { "id": "qari-markdown-mixed", "name": "QARI Markdown Mixed Dataset", "type": "dataset", "country": "SA", "org": "NAMAA-Space", "license": "apache-2.0", "modality": "vision", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/datasets/NAMAA-Space/QariOCR-v0.3-markdown-mixed-dataset", "paper": "https://arxiv.org/abs/2506.02295" }, "size": "10K-100K rows", "year": 2025, "dialects": [ "msa" ], "notes": "Training data for Qari OCR: Arabic page images paired with markdown including diacritics.", "metrics": { "downloads": 530, "likes": 13, "lastModified": "2025-06-10" } }, { "id": "hifzguide-muaalem-mini", "name": "hifzguide muaalem mini", "type": "asr", "country": "INTL", "org": "sysofwan", "license": "agpl-3.0", "modality": "speech", "tasks": [ "asr", "quran" ], "links": { "hf": "https://huggingface.co/sysofwan/hifzguide-muaalem-mini" }, "dialects": [ "classical" ], "year": 2026, "notes": "A size-distilled, teacher-initialised student of the Muaalem Quran phoneme-recognition model.", "base_model": [ "obadx/muaalem-model-v3_2" ], "metrics": { "downloads": 527, "likes": 0, "lastModified": "2026-09-22" } }, { "id": "dr-ai-v2", "name": "DR AI V2", "type": "llm", "country": "INTL", "org": "ehab215", "license": "gemma", "modality": "text", "tasks": [ "chat", "instruction-tuning", "vision" ], "links": { "hf": "https://huggingface.co/ehab215/DR-AI-V2" }, "dialects": [ "egy" ], "year": 2026, "notes": "A self-contained 4B conversational medical assistant for Egyptian Arabic and English.", "base_model": [ "google/medgemma-4b-it" ], "metrics": { "downloads": 519, "likes": 0, "lastModified": "2026-09-26" } }, { "id": "king-saud-university", "name": "King Saud University", "type": "org", "country": "SA", "org": "King Saud University", "license": "cc-by-nc-4.0", "modality": "none", "tasks": [ "research" ], "links": { "hf": "https://huggingface.co/faisalq/SaudiBERT" }, "notes": "SaudiBERT, Saudi dialect corpora (STMC, SFC)", "metrics": { "downloads": 518, "likes": 21, "lastModified": "2024-05-23" } }, { "id": "mmlu-arabic-fi", "name": "MMLU Arabic (FreedomIntelligence)", "type": "benchmark", "country": "INTL", "org": "FreedomIntelligence", "license": "mit", "modality": "text", "tasks": [ "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/FreedomIntelligence/MMLU_Arabic" }, "year": 2023, "dialects": [ "msa" ], "notes": "MMLU translated into Arabic by gpt-3.5-turbo.", "metrics": { "downloads": 518, "likes": 1, "lastModified": "2023-08-06" } }, { "id": "saudibert", "name": "SaudiBERT", "type": "llm", "country": "SA", "org": "King Saud University", "license": "cc-by-nc-4.0", "modality": "text", "tasks": [ "encoder" ], "links": { "hf": "https://huggingface.co/faisalq/SaudiBERT" }, "dialects": [ "gulf" ], "notes": "Saudi dialect BERT, trained on 141M tweets (STMC) + forums", "metrics": { "downloads": 518, "likes": 21, "lastModified": "2024-05-23" } }, { "id": "mkqa", "name": "MKQA", "type": "dataset", "country": "INTL", "org": "Apple", "license": "cc-by-sa-3.0", "modality": "text", "tasks": [ "qa" ], "links": { "github": "https://github.com/apple/ml-mkqa", "hf": "https://huggingface.co/datasets/apple/mkqa", "paper": "https://doi.org/10.1162/tacl_a_00433" }, "dialects": [ "mixed" ], "size": "10,000 sentences", "year": 2020, "tags": [ "multilingual" ], "notes": "10k question-answer pairs aligned across 26 languages, including Arabic.", "metrics": { "downloads": 516, "likes": 42, "lastModified": "2024-01-18" } }, { "id": "noor-platform-hadith", "name": "noor platform hadith", "type": "dataset", "country": "INTL", "org": "hozifa1", "license": "unknown", "modality": "text", "tasks": [ "hadith" ], "links": { "hf": "https://huggingface.co/datasets/hozifa1/noor-platform-hadith" }, "dialects": [ "classical" ], "year": 2026, "notes": "JSON hadith books and chapters (e.g., Abu Dawud) from the Noor platform.", "metrics": { "downloads": 515, "likes": 0, "lastModified": "2026-09-03" } }, { "id": "rightnow-arabic-0-5b-turbo", "name": "RightNow-Arabic-0.5B-Turbo", "type": "llm", "country": "INTL", "org": "RightNowAI", "license": "apache-2.0", "modality": "text", "tasks": [ "instruction-tuning", "pretraining" ], "links": { "hf": "https://huggingface.co/RightNowAI/RightNow-Arabic-0.5B-Turbo" }, "size": "0.5B", "on_device": true, "year": 2026, "notes": "RightNow-Arabic-0.5B-Turbo is a 518M-parameter Arabic-specialized language model built on top of Qwen2.5-0.5B via vocabulary injection.", "base_model": [ "qwen/qwen2.5-0.5b" ], "metrics": { "downloads": 513, "likes": 6, "lastModified": "2026-04-10" } }, { "id": "lahgtna-levantine-tts", "name": "Lahgtna Levantine TTS", "type": "dataset", "country": "EG", "org": "Mohammed Aly", "license": "cc-by-4.0", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/datasets/mohammedaly22/lahgtna-levantine-tts" }, "dialects": [ "lev" ], "notes": "Levantine Arabic TTS dataset from the Lahgtna project.", "metrics": { "downloads": 512, "likes": 4, "lastModified": "2026-05-30" } }, { "id": "alghafa-translated", "name": "AlGhafa Arabic LLM Benchmark (Translated)", "type": "benchmark", "country": "INTL", "org": "Open Arabic LLM Leaderboard (OALL)", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/OALL/AlGhafa-Arabic-LLM-Benchmark-Translated" }, "year": 2024, "dialects": [ "msa" ], "notes": "Machine-translated companion tasks to AlGhafa native, used in OALL v1.", "metrics": { "downloads": 511, "likes": 2, "lastModified": "2024-03-31" } }, { "id": "habibi-tts-doda-darija", "name": "habibi-tts-doda-darija", "type": "tts", "country": "MA", "org": "Jip7e", "license": "apache-2.0", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/Jip7e/habibi-tts-doda-darija" }, "year": 2026, "dialects": [ "magh" ], "notes": "Habibi TTS voice for Moroccan Darija.", "base_model": [ "swivid/habibi-tts" ], "metrics": { "downloads": 510, "likes": 1, "lastModified": "2026-08-29" } }, { "id": "bert-base-arabic-camelbert-mix-did-nadi", "name": "bert base arabic camelbert mix did nadi", "type": "llm", "country": "AE", "org": "CAMeL Lab, NYUAD", "license": "apache-2.0", "modality": "text", "tasks": [ "encoder", "dialect-id", "pretraining" ], "links": { "hf": "https://huggingface.co/CAMeL-Lab/bert-base-arabic-camelbert-mix-did-nadi" }, "dialects": [ "mixed" ], "year": 2021, "notes": "For the fine-tuning, we used the NADI Coountry-level dataset, which includes 21 labels.", "metrics": { "downloads": 505, "likes": 0, "lastModified": "2021-10-17" } }, { "id": "hadith-datasets", "name": "hadith datasets", "type": "dataset", "country": "INTL", "org": "meeAtif", "license": "mit", "modality": "text", "tasks": [ "hadith" ], "links": { "hf": "https://huggingface.co/datasets/meeAtif/hadith_datasets" }, "dialects": [ "classical" ], "size": "10K–100K rows", "year": 2026, "notes": "An open-source collection of authenticated Hadiths from the six major books of Sunnah, available in both JSON and CSV formats for research, study.", "metrics": { "downloads": 505, "likes": 14, "lastModified": "2026-02-09" } }, { "id": "islamic-scholars-prosopography", "name": "islamic scholars prosopography", "type": "dataset", "country": "INTL", "org": "hozifa1", "license": "cc-by-4.0", "modality": "text", "tasks": [ "qa" ], "links": { "hf": "https://huggingface.co/datasets/hozifa1/islamic-scholars-prosopography" }, "size": "10K–100K rows", "year": 2026, "notes": "مستودع موحد مفتوح المصدر يضم أضخم قواعد البيانات التراجمية الأكاديمية المهيكلة لعلماء وأعلام العالم الإسلامي وحواضر الأندلس والمشرق والمغرب، والمستخرجة من.", "metrics": { "downloads": 501, "likes": 0, "lastModified": "2026-09-07" } }, { "id": "arabic-pp-ocrv3-mobile-rec", "name": "arabic PP OCRv3 mobile rec", "type": "ocr", "country": "INTL", "org": "PaddlePaddle", "license": "apache-2.0", "modality": "vision", "tasks": [ "ocr", "vision" ], "links": { "hf": "https://huggingface.co/PaddlePaddle/arabic_PP-OCRv3_mobile_rec" }, "year": 2025, "notes": "arabicPP-OCRv3mobilerec is a text line recognition model within the PP-OCRv3rec series, developed by the PaddleOCR team.", "metrics": { "downloads": 500, "likes": 4, "lastModified": "2025-07-22" } }, { "id": "documents-egyptian-arabic", "name": "documents Egyptian Arabic", "type": "dataset", "country": "INTL", "org": "ISLAM-PO", "license": "apache-2.0", "modality": "text", "tasks": [ "translation", "asr", "qa" ], "links": { "hf": "https://huggingface.co/datasets/ISLAM-PO/documents-Egyptian-Arabic" }, "dialects": [ "egy" ], "size": "<1K rows", "year": 2026, "notes": "Source-specific configurations are explicitly declared in the dataset metadata; use the config that matches the schema", "metrics": { "downloads": 500, "likes": 2, "lastModified": "2026-09-25" } }, { "id": "qwen3-vl-2b-persian-arabic-ocr-v1-0", "name": "Qwen3-VL-2B-Persian-Arabic-Ocr-v1.0", "type": "ocr", "country": "INTL", "org": "Mohajesmaeili", "license": "apache-2.0", "modality": "vision", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/mohajesmaeili/Qwen3-VL-2B-Persian-Arabic-Ocr-v1.0" }, "size": "2.1B", "year": 2025, "on_device": true, "dialects": [ "msa" ], "notes": "Qwen3-VL 2B OCR for Persian and Arabic scripts.", "base_model": [ "qwen/qwen3-vl-2b-instruct" ], "metrics": { "downloads": 500, "likes": 20, "lastModified": "2025-12-23" } }, { "id": "masc-massive-arabic-speech-corpus", "name": "MASC: Massive Arabic Speech Corpus", "type": "dataset", "country": "INTL", "org": "Appswave", "license": "cc-by-4.0", "modality": "speech", "tasks": [ "asr" ], "links": { "website": "https://ieee-dataport.org/open-access/masc-massive-arabic-speech-corpus", "hf": "https://huggingface.co/datasets/pain/MASC", "paper": "https://ieeexplore.ieee.org/abstract/document/10022652" }, "dialects": [ "mixed" ], "size": "1,000 hours", "year": 2022, "notes": "This corpus is a dataset that contains 1,000 hours of speech sampled at 16kHz and crawled from over 700 YouTube channels.", "metrics": { "downloads": 498, "likes": 10, "lastModified": "2023-06-12" } }, { "id": "covid-19-disinformation", "name": "COVID 19 disinformation", "type": "dataset", "country": "QA", "org": "QCRI", "license": "cc-by-4.0", "modality": "text", "tasks": [ "nlp" ], "links": { "hf": "https://huggingface.co/datasets/QCRI/COVID-19-disinformation" }, "size": "10K–100K rows", "year": 2024, "notes": "This repository contains a multilingual dataset related to the COVID-19 infodemic, annotated with fine-grained labels.", "metrics": { "downloads": 495, "likes": 1, "lastModified": "2024-09-09" } }, { "id": "common-voice-18-arabic", "name": "common voice 18 arabic", "type": "dataset", "country": "INTL", "org": "MohamedRashad", "license": "cc", "modality": "speech", "tasks": [ "asr", "speech" ], "links": { "hf": "https://huggingface.co/datasets/MohamedRashad/common-voice-18-arabic" }, "size": "100K–1M rows", "year": 2025, "notes": "This dataset is an unofficial Arabic-only extraction of Mozilla Common Voice Corpus 18.0, prepared for Automatic Speech Recognition.", "metrics": { "downloads": 488, "likes": 5, "lastModified": "2025-12-27" } }, { "id": "arabic-legal-documents-ocr-1-0", "name": "arabic-legal-documents-ocr-1.0", "type": "ocr", "country": "EG", "org": "Bakrianoo", "license": "gemma", "modality": "vision", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/bakrianoo/arabic-legal-documents-ocr-1.0" }, "size": "4.3B", "year": 2026, "dialects": [ "msa" ], "notes": "Gemma-based OCR for Arabic legal documents.", "base_model": [ "google/gemma-3-4b-it" ], "metrics": { "downloads": 486, "likes": 48, "lastModified": "2026-02-04" } }, { "id": "zipformer-p-arabic-v3", "name": "zipformer_p-arabic-v3", "type": "asr", "country": "INTL", "org": "Quran Lab", "license": "other", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/Quran-Lab/zipformer_p-arabic-v3" }, "year": 2026, "dialects": [ "classical" ], "notes": "Zipformer ASR for Quranic Arabic.", "metrics": { "downloads": 474, "likes": 52, "lastModified": "2026-09-29" } }, { "id": "cohere-transcribe-arabic-07-2026-dialectal-v2", "name": "cohere-transcribe-arabic-07-2026-dialectal-v2", "type": "asr", "country": "EG", "org": "oddadmix", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/oddadmix/cohere-transcribe-arabic-07-2026-dialectal-v2" }, "size": "2.1B", "year": 2026, "dialects": [ "mixed" ], "notes": "Cohere Transcribe Arabic fine-tuned for dialectal speech.", "base_model": [ "coherelabs/cohere-transcribe-arabic-07-2026" ], "metrics": { "downloads": 473, "likes": 0, "lastModified": "2026-07-16" } }, { "id": "spokennativqa", "name": "SpokenNativQA", "type": "dataset", "country": "QA", "org": "QCRI", "license": "cc-by-nc-sa-4.0", "modality": "speech", "tasks": [ "qa", "asr" ], "links": { "hf": "https://huggingface.co/datasets/QCRI/SpokenNativQA" }, "size": "10K-100K rows", "year": 2025, "dialects": [ "mixed" ], "notes": "Spoken everyday queries with manually curated answers in multiple languages including Arabic.", "metrics": { "downloads": 473, "likes": 3, "lastModified": "2025-05-29" } }, { "id": "nile-chat-12b", "name": "Nile-Chat-12B", "type": "llm", "country": "AE", "org": "MBZUAI-Paris Lab", "license": "gemma", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/MBZUAI-Paris/Nile-Chat-12B" }, "base_model": [ "google/gemma-3-12b-pt" ], "size": "11.8B", "year": 2025, "dialects": [ "egy" ], "notes": "Nile-Chat 12B, Egyptian dialect model handling both Arabic and Arabizi script.", "metrics": { "downloads": 472, "likes": 15, "lastModified": "2025-07-10" } }, { "id": "baseer-nakba", "name": "Baseer Nakba", "type": "ocr", "country": "SA", "org": "Misraj AI", "license": "cc-by-nc-sa-4.0", "modality": "multimodal", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/Misraj/Baseer__Nakba", "paper": "https://arxiv.org/abs/2509.18174" }, "year": 2026, "notes": "Merged Baseer Arabic document OCR VLM variant on Qwen2.5-VL.", "base_model": [ "qwen2.5vl" ], "metrics": { "downloads": 469, "likes": 7, "lastModified": "2026-08-18" } }, { "id": "openthoughts3-en-ar-midtrain", "name": "openthoughts3 en ar midtrain", "type": "dataset", "country": "INTL", "org": "SultanR", "license": "unknown", "modality": "text", "tasks": [ "chat", "translation" ], "links": { "hf": "https://huggingface.co/datasets/SultanR/openthoughts3-en-ar-midtrain" }, "size": "1M–10M rows", "year": 2026, "notes": "Arabic translation of the OpenThoughts31.2M split of smoltalk2 (config Mid).", "metrics": { "downloads": 466, "likes": 0, "lastModified": "2026-08-17" } }, { "id": "nemotron-r1-en-ar-midtrain", "name": "nemotron r1 en ar midtrain", "type": "dataset", "country": "INTL", "org": "SultanR", "license": "unknown", "modality": "text", "tasks": [ "chat", "translation", "vision" ], "links": { "hf": "https://huggingface.co/datasets/SultanR/nemotron-r1-en-ar-midtrain" }, "size": "1M–10M rows", "year": 2026, "notes": "Arabic translation of the LlamaNemotronPostTrainingDatasetreasoningr1 split of smoltalk2 (config Mid, pinned revision fc6cc21).", "metrics": { "downloads": 463, "likes": 1, "lastModified": "2026-08-17" } }, { "id": "arabic-mpnet-base-all-nli-triplet", "name": "Arabic mpnet base all nli triplet", "type": "embedding", "country": "SA", "org": "Omartificial-Intelligence-Space", "license": "apache-2.0", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/Omartificial-Intelligence-Space/Arabic-mpnet-base-all-nli-triplet" }, "year": 2025, "notes": "This is a sentence-transformers model finetuned from tomaarsen/mpnet-base-all-nli-triplet on the Omartificial-Intelligence-Space/arabic-nli-triplet dataset.", "base_model": [ "tomaarsen/mpnet-base-all-nli-triplet" ], "metrics": { "downloads": 462, "likes": 10, "lastModified": "2025-01-23" } }, { "id": "cidar", "name": "CIDAR", "type": "dataset", "country": "SA", "org": "ARBML", "license": "apache-2.0", "modality": "text", "tasks": [ "nlp" ], "links": { "hf": "https://huggingface.co/datasets/arbml/CIDAR" }, "notes": "Culturally relevant instruction dataset (10K pairs)", "metrics": { "downloads": 460, "likes": 57, "lastModified": "2025-07-01" } }, { "id": "ar-sarcasm", "name": "ar sarcasm", "type": "dataset", "country": "INTL", "org": "iabufarha", "license": "mit", "modality": "text", "tasks": [ "sarcasm", "sentiment" ], "links": { "hf": "https://huggingface.co/datasets/iabufarha/ar_sarcasm" }, "size": "10K–100K rows", "year": 2024, "notes": "ArSarcasm: 10,547 tweets labelled for sarcasm, sentiment and dialect, built from SemEval 2017 and ASTD.", "metrics": { "downloads": 457, "likes": 18, "lastModified": "2024-01-09" } }, { "id": "silma-embedding-sts-v0-1", "name": "silma-embedding-sts-v0.1", "type": "embedding", "country": "INTL", "org": "SILMA AI", "license": "apache-2.0", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/silma-ai/silma-embedding-sts-v0.1" }, "size": "135M", "year": 2024, "on_device": true, "dialects": [ "msa" ], "notes": "SILMA embedding model tuned for Arabic semantic textual similarity.", "base_model": [ "silma-ai/silma-embeddding-matryoshka-0.1", "silma-ai/silma-embedding-matryoshka-v0.1" ], "metrics": { "downloads": 455, "likes": 6, "lastModified": "2024-11-04" } }, { "id": "wav2vec2-xls-r-300m-iqraeval", "name": "wav2vec2 xls r 300m iqraeval", "type": "asr", "country": "YE", "org": "Fatimah Emad Eldin", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr", "vision", "quran" ], "links": { "hf": "https://huggingface.co/FatimahEmadEldin/wav2vec2-xls-r-300m-iqraeval" }, "dialects": [ "classical", "msa" ], "size": "300M", "on_device": true, "year": 2026, "notes": "This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m specifically adapted for the Iqra'Eval 2026 Shared Task.", "metrics": { "downloads": 454, "likes": 3, "lastModified": "2026-03-01" } }, { "id": "madinahqa", "name": "MadinahQA", "type": "benchmark", "country": "AE", "org": "MBZUAI", "license": "cc-by-nc-4.0", "modality": "text", "tasks": [ "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/MBZUAI/MadinahQA" }, "year": 2024, "dialects": [ "msa" ], "notes": "Arabic language and grammar multiple-choice QA split released with ArabicMMLU.", "metrics": { "downloads": 453, "likes": 1, "lastModified": "2024-09-17" } }, { "id": "arcd", "name": "ARCD", "type": "benchmark", "country": "INTL", "org": "Mozannar et al.", "license": "mit", "modality": "text", "tasks": [ "qa", "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/hsseinmz/arcd" }, "size": "1.4k questions", "year": 2019, "dialects": [ "msa" ], "notes": "Arabic Reading Comprehension Dataset of 1,395 Wikipedia-based questions from the SOQAL paper.", "metrics": { "downloads": 452, "likes": 12, "lastModified": "2024-01-09" } }, { "id": "common-voice-arabic-12-0-augmented", "name": "common voice Arabic 12.0 Augmented", "type": "dataset", "country": "INTL", "org": "Salama1429", "license": "unknown", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/datasets/Salama1429/common_voice_Arabic_12.0_Augmented" }, "size": "10K–100K rows", "year": 2022, "notes": "Augmented Common Voice 12.0 Arabic: 63,546 training audio-sentence pairs at 16 kHz.", "metrics": { "downloads": 449, "likes": 2, "lastModified": "2022-12-24" } }, { "id": "imageeval-arabicnlp26", "name": "ImageEval-ArabicNLP26", "type": "benchmark", "country": "QA", "org": "QCRI", "license": "cc-by-nc-sa-4.0", "modality": "multimodal", "tasks": [ "qa", "evaluation", "speech", "vision" ], "links": { "hf": "https://huggingface.co/datasets/QCRI/ImageEval-ArabicNLP26" }, "size": "10K–100K rows", "year": 2026, "notes": "It covers both of the shared task's tasks: AynVQA (Task 1), a culturally grounded Arabic multimodal benchmark for spoken visual question answering.", "metrics": { "downloads": 449, "likes": 4, "lastModified": "2026-08-30" } }, { "id": "quranic-audio-nonnative", "name": "Quranic Audio (non-Arabic speakers)", "type": "dataset", "country": "INTL", "org": "RetaSy", "license": "unknown", "modality": "speech", "tasks": [ "asr", "pronunciation" ], "links": { "hf": "https://huggingface.co/datasets/RetaSy/quranic_audio_dataset" }, "year": 2024, "dialects": [ "classical" ], "notes": "Crowdsourced and labeled Quran recitations from non-Arabic speakers.", "metrics": { "downloads": 448, "likes": 12, "lastModified": "2024-05-14" } }, { "id": "arabic-vlm-full-pearl", "name": "Arabic-VLM-Full-Pearl", "type": "dataset", "country": "INTL", "org": "MohamedRashad", "license": "unknown", "modality": "vision", "tasks": [ "multimodal" ], "links": { "hf": "https://huggingface.co/datasets/MohamedRashad/Arabic-VLM-Full-Pearl" }, "notes": "309K multimodal examples for VLM training", "metrics": { "downloads": 443, "likes": 10, "lastModified": "2025-12-08" } }, { "id": "mmbert-base-arabic-nli", "name": "mmbert base arabic nli", "type": "embedding", "country": "SA", "org": "Omartificial-Intelligence-Space", "license": "apache-2.0", "modality": "text", "tasks": [ "embedding", "nli" ], "links": { "hf": "https://huggingface.co/Omartificial-Intelligence-Space/mmbert-base-arabic-nli" }, "year": 2025, "notes": "Arabic sentence embedding model fine-tuned from mmBERT-base on 900K NLI triplets with GISTEmbedLoss (768-dim).", "base_model": [ "jhu-clsp/mmbert-base" ], "metrics": { "downloads": 442, "likes": 1, "lastModified": "2025-10-08" } }, { "id": "hadra-tts-f5", "name": "Hadra-TTS-f5", "type": "tts", "country": "DZ", "org": "Algerian NLP", "license": "apache-2.0", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/algerian-nlp/Hadra-TTS-f5" }, "year": 2026, "dialects": [ "magh" ], "notes": "F5-based TTS for Algerian Darija; Algeria noted.", "base_model": [ "ibrahimsalah/arabic-f5-tts-v2" ], "metrics": { "downloads": 441, "likes": 0, "lastModified": "2026-09-23" } }, { "id": "silma-tts", "name": "SILMA TTS v1", "type": "tts", "country": "INTL", "org": "SILMA AI", "license": "apache-2.0", "modality": "speech", "tasks": [ "tts", "voice-cloning" ], "links": { "github": "https://github.com/SILMA-AI/silma-tts", "hf": "https://huggingface.co/silma-ai/silma-tts" }, "year": 2026, "notes": "Open 150M bilingual Arabic/English TTS on the F5-TTS architecture; voice cloning from 8 s of audio.", "size": "150M", "on_device": true, "metrics": { "downloads": 437, "likes": 35, "lastModified": "2026-08-23" } }, { "id": "aradice-arabicmmlu-lev", "name": "AraDICE-ArabicMMLU-lev", "type": "benchmark", "country": "QA", "org": "QCRI", "license": "cc-by-nc-sa-4.0", "modality": "text", "tasks": [ "qa", "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/QCRI/AraDICE-ArabicMMLU-lev" }, "dialects": [ "lev" ], "size": "10K–100K rows", "year": 2024, "notes": "The AraDiCE dataset is crafted to assess the dialectal and cultural understanding of large language models (LLMs) within Arabic-speaking contexts.", "metrics": { "downloads": 425, "likes": 0, "lastModified": "2024-11-08" } }, { "id": "gate-reranker-v1", "name": "GATE-Reranker-V1", "type": "embedding", "country": "SA", "org": "NAMAA-Space", "license": "apache-2.0", "modality": "text", "tasks": [ "reranking" ], "links": { "hf": "https://huggingface.co/NAMAA-Space/GATE-Reranker-V1" }, "size": "135M", "year": 2024, "on_device": true, "dialects": [ "msa" ], "notes": "Arabic reranker from NAMAA built on GATE-AraBERT.", "base_model": [ "omartificial-intelligence-space/gate-arabert-v1" ], "metrics": { "downloads": 423, "likes": 11, "lastModified": "2025-04-03" } }, { "id": "arabic-speech-corpus", "name": "Arabic Speech Corpus", "type": "dataset", "country": "INTL", "org": "halabi2016", "license": "cc-by-4.0", "modality": "speech", "tasks": [ "tts" ], "links": { "hf": "https://huggingface.co/datasets/halabi2016/arabic_speech_corpus" }, "notes": "Nawar Halabi's Levantine Arabic TTS corpus (CC-BY-4.0)", "metrics": { "downloads": 417, "likes": 38, "lastModified": "2024-08-14" } }, { "id": "msa-omnivoice-tts-v1", "name": "msa omnivoice tts v1", "type": "dataset", "country": "EG", "org": "oddadmix", "license": "unknown", "modality": "speech", "tasks": [ "asr", "tts", "speech" ], "links": { "hf": "https://huggingface.co/datasets/oddadmix/msa-omnivoice-tts-v1" }, "dialects": [ "msa" ], "size": "10K–100K rows", "year": 2026, "notes": "It is intended for training and fine-tuning Arabic speech models, including Text-to-Speech (TTS), Automatic Speech Recognition.", "metrics": { "downloads": 412, "likes": 1, "lastModified": "2026-07-08" } }, { "id": "al-kawakib-magazine-ocr", "name": "al kawakib magazine ocr", "type": "dataset", "country": "INTL", "org": "amrosama", "license": "cc-by-4.0", "modality": "multimodal", "tasks": [ "ocr", "instruction-tuning", "vision" ], "links": { "hf": "https://huggingface.co/datasets/amrosama/al-kawakib-magazine-ocr" }, "size": "10K–100K rows", "year": 2026, "notes": "This dataset contains rendered Arabic magazine page images paired with page-level text and line-level bounding boxes.", "metrics": { "downloads": 411, "likes": 3, "lastModified": "2026-07-25" } }, { "id": "ajgt-twitter-ar", "name": "ajgt twitter ar", "type": "dataset", "country": "INTL", "org": "komari6", "license": "['unknown']", "modality": "text", "tasks": [ "sentiment" ], "links": { "hf": "https://huggingface.co/datasets/komari6/ajgt_twitter_ar" }, "size": "1K–10K rows", "year": 2024, "notes": "Arabic Jordanian General Tweets: 1,800 tweets labelled positive or negative, in MSA or Jordanian dialect.", "dialects": [ "msa", "lev" ], "metrics": { "downloads": 410, "likes": 4, "lastModified": "2024-01-09" } }, { "id": "bge-m3-law", "name": "bge m3 law", "type": "embedding", "country": "INTL", "org": "mhaseeb1604", "license": "unknown", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/mhaseeb1604/bge-m3-law" }, "year": 2024, "notes": "This model is a fine-tuned version of the BAAI/bge-m3 model, which is specialized for sentence similarity tasks in Arabic legal texts in both Arabic.", "base_model": [ "baai/bge-m3" ], "metrics": { "downloads": 404, "likes": 6, "lastModified": "2024-10-09" } }, { "id": "labr", "name": "LABR", "type": "dataset", "country": "EG", "org": "Cairo University", "license": "gpl-2.0", "modality": "text", "tasks": [ "sentiment" ], "links": { "github": "https://github.com/mohamedadaly/LABR", "hf": "https://huggingface.co/datasets/mohamedadaly/labr", "paper": "https://aclanthology.org/P13-2088.pdf" }, "dialects": [ "mixed" ], "size": "63,257 sentences", "year": 2013, "notes": "The largest sentiment analysis dataset to-date for the Arabic language.", "metrics": { "downloads": 402, "likes": 3, "lastModified": "2024-08-08" } }, { "id": "whisper-tunisian-dialect", "name": "whisper-tunisian-dialect", "type": "asr", "country": "TN", "org": "TuniSpeech-AI", "license": "cc-by-nc-4.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/TuniSpeech-AI/whisper-tunisian-dialect" }, "size": "1.5B", "year": 2026, "dialects": [ "magh" ], "notes": "Whisper fine-tuned on Tunisian dialect.", "metrics": { "downloads": 402, "likes": 4, "lastModified": "2026-02-13" } }, { "id": "moulsot-full", "name": "MoulSot-Full", "type": "dataset", "country": "MA", "org": "atlasia", "license": "unknown", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/datasets/atlasia/MoulSot-Full" }, "size": "1500h", "year": 2025, "dialects": [ "magh" ], "notes": "Large Moroccan Darija speech corpus of about 1,500 hours.", "metrics": { "downloads": 399, "likes": 12, "lastModified": "2026-05-05" } }, { "id": "bert-large-arabertv02-twitter", "name": "bert large arabertv02 twitter", "type": "llm", "country": "LB", "org": "AUB MIND Lab", "license": "unknown", "modality": "text", "tasks": [ "encoder", "pretraining" ], "links": { "hf": "https://huggingface.co/aubmindlab/bert-large-arabertv02-twitter" }, "dialects": [ "mixed" ], "year": 2023, "notes": "AraBERTv0.2-Twitter-base/large are two new models for Arabic dialects and tweets.", "metrics": { "downloads": 394, "likes": 4, "lastModified": "2023-04-26" } }, { "id": "shifaa-medical", "name": "Shifaa Medical", "type": "dataset", "country": "EG", "org": "Ahmed Selem", "license": "apache-2.0", "modality": "text", "tasks": [ "nlp" ], "links": { "hf": "https://huggingface.co/datasets/Ahmed-Selem/Shifaa_Arabic_Medical_Consultations" }, "notes": "Arabic medical consultation dataset", "metrics": { "downloads": 394, "likes": 13, "lastModified": "2025-02-28" } }, { "id": "arabicweb24", "name": "ArabicWeb24", "type": "dataset", "country": "INTL", "org": "lightonai", "license": "odc-by", "modality": "text", "tasks": [ "nlp" ], "links": { "hf": "https://huggingface.co/datasets/lightonai/ArabicWeb24" }, "notes": "39B+ tokens of high-quality Arabic web content", "metrics": { "downloads": 392, "likes": 24, "lastModified": "2024-09-23" } }, { "id": "jabarti-llm-dataset", "name": "jabarti llm dataset", "type": "dataset", "country": "EG", "org": "Bakrianoo", "license": "cc-by-sa-4.0", "modality": "text", "tasks": [ "qa", "pretraining" ], "links": { "hf": "https://huggingface.co/datasets/bakrianoo/jabarti-llm-dataset" }, "dialects": [ "egy" ], "size": "1M–10M rows", "year": 2026, "notes": "Cleaned, section-chunked training corpus for a small bilingual LLM.", "metrics": { "downloads": 390, "likes": 24, "lastModified": "2026-09-23" } }, { "id": "arabic-ocr-synthetic-scans-faker-300k", "name": "arabic ocr synthetic scans faker 300k", "type": "dataset", "country": "INTL", "org": "loay", "license": "cc-by-4.0", "modality": "multimodal", "tasks": [ "ocr", "vision" ], "links": { "hf": "https://huggingface.co/datasets/loay/arabic-ocr-synthetic-scans-faker-300k" }, "size": "100K–1M rows", "year": 2026, "notes": "A large-scale synthetic dataset of 300,000 Arabic book pages generated to mimic real-world scanning imperfections.", "metrics": { "downloads": 387, "likes": 7, "lastModified": "2026-02-12" } }, { "id": "nawah-50m-rag-support-2k", "name": "Nawah 50M RAG Support 2K", "type": "llm", "country": "EG", "org": "oddadmix", "license": "apache-2.0", "modality": "text", "tasks": [ "chat", "embedding", "qa" ], "links": { "hf": "https://huggingface.co/oddadmix/Nawah-50M-RAG-Support-2K" }, "dialects": [ "msa" ], "size": "50M", "on_device": true, "year": 2026, "notes": "A 51.8M-parameter Arabic (MSA) retrieval-augmented customer-support answerer, trained from scratch.", "base_model": [ "oddadmix/50m-2048-emhotob" ], "metrics": { "downloads": 387, "likes": 5, "lastModified": "2026-08-19" } }, { "id": "arabic-marbert-sentiment", "name": "arabic MARBERT sentiment", "type": "llm", "country": "INTL", "org": "Ammar-alhaj-ali", "license": "unknown", "modality": "text", "tasks": [ "sentiment", "classification" ], "links": { "hf": "https://huggingface.co/Ammar-alhaj-ali/arabic-MARBERT-sentiment" }, "year": 2022, "notes": "MARBERT fine-tuned for three-class Arabic sentiment analysis (positive, negative, neutral).", "metrics": { "downloads": 385, "likes": 11, "lastModified": "2022-08-09" } }, { "id": "laion-coco-5m-arabic-labels", "name": "laion coco 5m arabic labels", "type": "dataset", "country": "EG", "org": "oddadmix", "license": "cc-by-sa-4.0", "modality": "vision", "tasks": [ "ocr", "translation", "vision" ], "links": { "hf": "https://huggingface.co/datasets/oddadmix/laion-coco-5m-arabic-labels" }, "size": "1M–10M rows", "year": 2026, "notes": "An Arabic-labelled, image-bearing derivative of CaptionEmporium/laion-coco-13m-molmo-d-7b.", "metrics": { "downloads": 385, "likes": 0, "lastModified": "2026-08-29" } }, { "id": "darija-english-doda", "name": "Darija-English (DODa)", "type": "dataset", "country": "MA", "org": "Darija Open Dataset", "license": "cc", "modality": "text", "tasks": [ "translation" ], "links": { "hf": "https://huggingface.co/datasets/imomayiz/darija-english" }, "year": 2024, "dialects": [ "magh" ], "notes": "Darija-English pairs from the Darija Open Dataset (DODa) project.", "metrics": { "downloads": 384, "likes": 13, "lastModified": "2024-04-17" } }, { "id": "kfupm-arabic-ai-text-detection", "name": "KFUPM Arabic AI Text Detection", "type": "dataset", "country": "SA", "org": "KFUPM-JRCAI", "license": "unknown", "modality": "text", "tasks": [ "nlp" ], "links": { "hf": "https://huggingface.co/datasets/KFUPM-JRCAI/arabic-generated-abstracts" }, "notes": "Machine-generated Arabic text across multiple LLMs", "metrics": { "downloads": 383, "likes": 4, "lastModified": "2026-05-22" } }, { "id": "neoarabert", "name": "NeoAraBERT", "type": "llm", "country": "INTL", "org": "U4RASD (ACRPS)", "license": "cc-by-sa-4.0", "modality": "text", "tasks": [ "encoder", "embedding" ], "links": { "hf": "https://huggingface.co/U4RASD/NeoAraBERT" }, "dialects": [ "classical", "msa" ], "year": 2026, "notes": "NeoAraBERT_Mix: NeoBERT-based Arabic encoder from ACRPS U4RASD and AUB, ranking first on 18 of 23 tasks.", "base_model": [ "u4rasd/neoarabert" ], "metrics": { "downloads": 382, "likes": 10, "lastModified": "2026-09-08" } }, { "id": "whisper-medium-darija", "name": "whisper-medium-darija", "type": "asr", "country": "MA", "org": "Youssef Chafiqui", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/ychafiqui/whisper-medium-darija" }, "size": "764M", "year": 2024, "on_device": true, "dialects": [ "magh" ], "notes": "Whisper medium fine-tuned on Moroccan Darija.", "base_model": [ "openai/whisper-medium" ], "metrics": { "downloads": 381, "likes": 7, "lastModified": "2024-09-27" } }, { "id": "waqf-ocr-hand-written-v1", "name": "waqf ocr hand written v1", "type": "ocr", "country": "EG", "org": "Waqf AI", "license": "apache-2.0", "modality": "vision", "tasks": [ "chat", "ocr", "evaluation", "vision" ], "links": { "hf": "https://huggingface.co/Waqf-AI/waqf-ocr-hand-written-v1" }, "year": 2026, "notes": "Fine-tuned from PaddlePaddle/PaddleOCR-VL-1.6 on khatt-augmented-arabic-20k-s3.", "metrics": { "downloads": 380, "likes": 5, "lastModified": "2026-06-09" } }, { "id": "wav2veclarge-quran-syllables-recognition", "name": "Wav2vecLarge_quran_syllables_recognition", "type": "asr", "country": "INTL", "org": "IbrahimSalah", "license": "unknown", "modality": "speech", "tasks": [ "asr", "quran" ], "links": { "hf": "https://huggingface.co/IbrahimSalah/Wav2vecLarge_quran_syllables_recognition" }, "dialects": [ "classical" ], "year": 2025, "notes": "This is fine tuned wav2vec2 model to recognize quran syllables from speech.", "metrics": { "downloads": 378, "likes": 6, "lastModified": "2025-01-05" } }, { "id": "rasam-1", "name": "RASAM 1", "type": "dataset", "country": "INTL", "org": "calfa-ai", "license": "apache-2.0", "modality": "multimodal", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/datasets/calfa-ai/RASAM-1" }, "size": "1K–10K rows", "year": 2026, "tags": [ "variants:1" ], "notes": "RASAM 1 is a specialized dataset for Handwritten Text Recognit (also: 1 variants)", "metrics": { "downloads": 375, "likes": 3, "lastModified": "2026-09-11" } }, { "id": "fineweb-edu-arabic", "name": "FineWeb-Edu Arabic", "type": "dataset", "country": "INTL", "org": "SultanR", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/datasets/SultanR/fineweb-edu-arabic" }, "size": "10M-100M rows", "year": 2026, "dialects": [ "msa" ], "notes": "Arabic translation of the FineWeb-Edu sample filtered by language score.", "metrics": { "downloads": 373, "likes": 1, "lastModified": "2026-08-17" } }, { "id": "jais-13b-chat", "name": "jais 13b chat", "type": "llm", "country": "AE", "org": "Inception AI (G42)", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/inception42/jais-13b-chat" }, "base_model": [ "inception-mbzuai/jais-13b" ], "size": "13B", "on_device": false, "year": 2024, "tags": [ "variants:1" ], "notes": "Instruction-tuned chat version of Jais-13B, a bilingual Arabic-English LLM (arXiv 2308.16149).", "metrics": { "downloads": 373, "likes": 168, "lastModified": "2024-09-11" } }, { "id": "arabic-morocco-speech-to-text", "name": "Arabic Morocco Speech To Text", "type": "asr", "country": "INTL", "org": "smerchi", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/smerchi/Arabic-Morocco-Speech_To_Text" }, "year": 2024, "notes": "Whisper large-v3 fine-tuned on Cleverlytics voice data for Moroccan Arabic speech recognition.", "dialects": [ "magh" ], "base_model": [ "openai/whisper-large-v3" ], "metrics": { "downloads": 372, "likes": 17, "lastModified": "2024-04-02" } }, { "id": "aramodernbert-base-sts", "name": "AraModernBert-Base-STS", "type": "embedding", "country": "SA", "org": "NAMAA-Space", "license": "apache-2.0", "modality": "text", "tasks": [ "chat", "embedding" ], "links": { "hf": "https://huggingface.co/NAMAA-Space/AraModernBert-Base-STS" }, "year": 2025, "notes": "This SentenceTransformer is fine-tuned from NAMAA-Space/AraModernBert-Base-V1.0, bringing strong arabic embeddings useful for a multiple.", "base_model": [ "namaa-space/aramodernbert-base-v1.0" ], "metrics": { "downloads": 371, "likes": 7, "lastModified": "2025-03-09" } }, { "id": "star-dataset-instructions", "name": "star dataset instructions", "type": "dataset", "country": "SA", "org": "KFUPM-JRCAI", "license": "unknown", "modality": "text", "tasks": [ "ner", "translation", "summarization", "qa", "embedding", "instruction-tuning" ], "links": { "hf": "https://huggingface.co/datasets/KFUPM-JRCAI/star-dataset-instructions" }, "size": "10M–100M rows", "year": 2026, "notes": "STAR Instructions is a large-scale Arabic instruction-tuning dataset built by rendering the 355 STAR Jinja2 prompt templates against their 87 source.", "metrics": { "downloads": 371, "likes": 1, "lastModified": "2026-09-03" } }, { "id": "arabic-tashkeel-flan-t5-small", "name": "arabic tashkeel flan t5 small", "type": "llm", "country": "INTL", "org": "Abdou", "license": "unknown", "modality": "text", "tasks": [ "tashkeel" ], "links": { "hf": "https://huggingface.co/Abdou/arabic-tashkeel-flan-t5-small" }, "year": 2024, "notes": "Flan-T5-small model that adds tashkeel (Arabic diacritics) to Arabic text.", "on_device": true, "metrics": { "downloads": 370, "likes": 4, "lastModified": "2024-10-22" } }, { "id": "blip3-grounding-1m-arabic", "name": "blip3 grounding 1m arabic", "type": "dataset", "country": "EG", "org": "oddadmix", "license": "apache-2.0", "modality": "multimodal", "tasks": [ "ocr", "translation", "vision" ], "links": { "hf": "https://huggingface.co/datasets/oddadmix/blip3-grounding-1m-arabic" }, "size": "1M–10M rows", "year": 2026, "notes": "Arabic version of the first 1,000,000 usable rows of Salesforce/blip3-grounding-50m.", "metrics": { "downloads": 365, "likes": 0, "lastModified": "2026-08-23" } }, { "id": "arabicner-wojood", "name": "ArabicNER-Wojood", "type": "llm", "country": "PS", "org": "SinaLab", "license": "mit", "modality": "text", "tasks": [ "ner" ], "links": { "hf": "https://huggingface.co/SinaLab/ArabicNER-Wojood" }, "year": 2023, "dialects": [ "msa" ], "notes": "Arabic nested NER model trained on Wojood; Palestine.", "metrics": { "downloads": 363, "likes": 10, "lastModified": "2024-03-20" } }, { "id": "hadra-asr-whisper-medium", "name": "Hadra-ASR-whisper-medium", "type": "asr", "country": "DZ", "org": "Algerian NLP", "license": "mit", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/algerian-nlp/Hadra-ASR-whisper-medium" }, "year": 2026, "dialects": [ "magh" ], "notes": "Whisper medium for Algerian Darija; Algeria noted.", "base_model": [ "openai/whisper-medium" ], "metrics": { "downloads": 362, "likes": 1, "lastModified": "2026-09-17" } }, { "id": "nilechat-3b", "name": "NileChat-3B", "type": "llm", "country": "INTL", "org": "UBC-NLP", "license": "qwen-research", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/UBC-NLP/NileChat-3B" }, "base_model": [ "qwen/qwen2.5-3b" ], "size": "3B", "on_device": true, "dialects": [ "egy", "magh" ], "notes": "Egyptian and Moroccan dialects", "metrics": { "downloads": 361, "likes": 23, "lastModified": "2026-05-11" } }, { "id": "visper", "name": "visper", "type": "dataset", "country": "AE", "org": "TII (UAE)", "license": "cc-by-nc-2.0", "modality": "speech", "tasks": [ "asr", "speech", "vision" ], "links": { "hf": "https://huggingface.co/datasets/tiiuae/visper" }, "size": "10K–100K rows", "year": 2025, "notes": "This repository contains ViSpeR, a large-scale dataset and models for Visual Speech Recognition for Arabic, Chinese, French, Arabic and Spanish.", "metrics": { "downloads": 360, "likes": 6, "lastModified": "2025-04-17" } }, { "id": "quranmb-v2", "name": "QuranMB.v2", "type": "benchmark", "country": "INTL", "org": "IqraEval", "license": "unknown", "modality": "speech", "tasks": [ "asr", "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/IqraEval/QuranMB.v2" }, "year": 2025, "dialects": [ "classical" ], "notes": "IqraEval benchmark for Quranic mispronunciation detection.", "metrics": { "downloads": 350, "likes": 3, "lastModified": "2025-07-05" } }, { "id": "egyptian-arabic-lectures", "name": "Egyptian Arabic Lectures", "type": "dataset", "country": "INTL", "org": "ismaeeelxd", "license": "mit", "modality": "speech", "tasks": [ "asr", "evaluation", "speech" ], "links": { "hf": "https://huggingface.co/datasets/ismaeeelxd/Egyptian-Arabic-Lectures" }, "dialects": [ "egy" ], "size": "1K–10K rows", "year": 2026, "notes": "The Egyptian Arabic Lectures dataset is a collection of transcribed audio clips.", "metrics": { "downloads": 347, "likes": 3, "lastModified": "2026-06-21" } }, { "id": "tafsir-mcp-data", "name": "tafsir mcp data", "type": "dataset", "country": "INTL", "org": "Tafsir Center", "license": "cc-by-4.0", "modality": "text", "tasks": [ "quran" ], "links": { "hf": "https://huggingface.co/datasets/tafsircenter/tafsir-mcp-data" }, "dialects": [ "classical" ], "size": "1K–10K rows", "year": 2026, "notes": "Quran and tafsir database behind tafsir-mcp, an MCP server giving AI assistants offline access to Quranic scholarship.", "metrics": { "downloads": 346, "likes": 4, "lastModified": "2026-05-13" } }, { "id": "khatt-v1-0-dataset", "name": "KHATT v1.0 dataset", "type": "dataset", "country": "INTL", "org": "johnlockejrr", "license": "mit", "modality": "vision", "tasks": [ "ocr", "htr" ], "links": { "hf": "https://huggingface.co/datasets/johnlockejrr/KHATT_v1.0_dataset" }, "year": 2024, "notes": "Line-level KHATT database of unconstrained handwritten Arabic text from 1000 writers, developed by a KFUPM research group.", "metrics": { "downloads": 345, "likes": 4, "lastModified": "2024-07-01" } }, { "id": "open-arabicaqa", "name": "Open-ArabicaQA", "type": "dataset", "country": "INTL", "org": "DataScienceUIBK", "license": "mit", "modality": "text", "tasks": [ "qa", "retrieval" ], "links": { "hf": "https://huggingface.co/datasets/abdoelsayed/Open-ArabicaQA" }, "size": "10K-100K rows", "year": 2024, "dialects": [ "msa" ], "notes": "Open-domain QA and retrieval splits of ArabicaQA on Hugging Face.", "metrics": { "downloads": 345, "likes": 10, "lastModified": "2024-03-27" } }, { "id": "aramix-translation-scores", "name": "AraMix-Translation-Scores", "type": "dataset", "country": "INTL", "org": "SultanR", "license": "odc-by", "modality": "text", "tasks": [ "translation" ], "links": { "hf": "https://huggingface.co/datasets/SultanR/AraMix-Translation-Scores" }, "size": "100M–1B rows", "year": 2026, "notes": "AdaMLLab/AraMix (minhashdeduped subset, 178,883,241 rows) with a machine-translation-detection score added to every document.", "metrics": { "downloads": 344, "likes": 0, "lastModified": "2026-07-18" } }, { "id": "algerian-darja-corpus", "name": "Algerian Darja Corpus", "type": "dataset", "country": "DZ", "org": "Touati Kamel", "license": "cc-by-4.0", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/datasets/touati-kamel/algerian-darja-corpus" }, "size": "10K-100K rows", "year": 2026, "dialects": [ "magh" ], "notes": "Algerian Darja text corpus with Zenodo DOI (Algeria).", "metrics": { "downloads": 343, "likes": 6, "lastModified": "2026-09-18" } }, { "id": "t5-arabic-summarization", "name": "t5-arabic-summarization", "type": "llm", "country": "INTL", "org": "malmarjeh", "license": "unknown", "modality": "text", "tasks": [ "task-specific" ], "links": { "hf": "https://huggingface.co/malmarjeh/t5-arabic-text-summarization" }, "notes": "Summarization - T5 for Arabic news summarization", "metrics": { "downloads": 343, "likes": 14, "lastModified": "2023-07-01" } }, { "id": "aramix-native", "name": "AraMix-Native", "type": "dataset", "country": "INTL", "org": "SultanR", "license": "odc-by", "modality": "text", "tasks": [ "translation", "diacritization" ], "links": { "hf": "https://huggingface.co/datasets/SultanR/AraMix-Native" }, "size": "100M–1B rows", "year": 2026, "notes": "A native-Arabic-filtered version of AdaMLLab/AraMix (minhashdeduped), derived from SultanR/AraMix-Translation-Scores.", "metrics": { "downloads": 341, "likes": 0, "lastModified": "2026-07-19" } }, { "id": "openresearcher-corpus-ar", "name": "openresearcher corpus ar", "type": "dataset", "country": "INTL", "org": "SultanR", "license": "unknown", "modality": "text", "tasks": [ "translation", "pretraining" ], "links": { "hf": "https://huggingface.co/datasets/SultanR/openresearcher-corpus-ar" }, "size": "1M–10M rows", "year": 2026, "notes": "Arabic translation of OpenResearcher-Corpus, an 11B token curated web corpus.", "metrics": { "downloads": 341, "likes": 0, "lastModified": "2026-08-17" } }, { "id": "ar-musa", "name": "Ar MUSA", "type": "dataset", "country": "INTL", "org": "Skhaled", "license": "afl-3.0", "modality": "multimodal", "tasks": [ "speech", "sentiment" ], "links": { "hf": "https://huggingface.co/datasets/Skhaled/Ar-MUSA" }, "size": "1K–10K rows", "year": 2025, "notes": "The Ar-MUSA directory contains annotated datasets organized by batches and annotation teams.", "metrics": { "downloads": 338, "likes": 3, "lastModified": "2025-04-14" } }, { "id": "aragen", "name": "AraGen", "type": "benchmark", "country": "AE", "org": "Inception AI (G42)", "license": "unknown", "modality": "text", "tasks": [ "evaluation" ], "links": { "hf": "https://huggingface.co/datasets/inception42/AraGen" }, "year": 2024, "notes": "Generative Arabic LLM benchmark with 3C3H scoring and dynamic periodic refresh, with MBZUAI.", "metrics": { "downloads": 337, "likes": 2, "lastModified": "2025-12-15" } }, { "id": "arabic-qwen3-5-ocr-v4", "name": "Arabic-Qwen3.5-OCR-v4", "type": "ocr", "country": "INTL", "org": "sherif1313", "license": "apache-2.0", "modality": "vision", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/sherif1313/Arabic-Qwen3.5-OCR-v4" }, "size": "853M", "year": 2026, "on_device": true, "dialects": [ "msa" ], "notes": "Qwen3.5-based Arabic OCR model.", "base_model": [ "qwen/qwen3.5-0.8b" ], "metrics": { "downloads": 336, "likes": 10, "lastModified": "2026-03-24" } }, { "id": "arabic-audio-collection-moroccan-ameed", "name": "arabic audio collection moroccan ameed", "type": "dataset", "country": "EG", "org": "oddadmix", "license": "other", "modality": "speech", "tasks": [ "tts", "asr", "diacritization", "sentiment", "speech" ], "links": { "hf": "https://huggingface.co/datasets/oddadmix/arabic-audio-collection-moroccan-ameed" }, "dialects": [ "magh" ], "size": "10K–100K rows", "year": 2026, "notes": "The Ameed Moroccan Arabic Speech Dataset is a large-scale, first-of-its-kind Arabic speech corpus containing approximately 176 hours of speech recordings.", "metrics": { "downloads": 335, "likes": 1, "lastModified": "2026-06-30" } }, { "id": "tunbert", "name": "TunBERT", "type": "llm", "country": "TN", "org": "InstaDeep / iCompass", "license": "mit", "modality": "text", "tasks": [ "embedding" ], "links": { "hf": "https://huggingface.co/tunis-ai/TunBERT" }, "size": "110M", "year": 2021, "dialects": [ "magh" ], "notes": "First BERT for Tunisian dialect (Derja), from InstaDeep and iCompass.", "metrics": { "downloads": 335, "likes": 15, "lastModified": "2025-09-29" } }, { "id": "arabiangpt", "name": "ArabianGPT", "type": "llm", "country": "SA", "org": "Prince Sultan University", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/riotu-lab/ArabianGPT-01B" }, "size": "0.1B", "notes": "GPT-2 for Arabic, RIOTU Lab", "on_device": true, "metrics": { "downloads": 334, "likes": 13, "lastModified": "2024-02-27" } }, { "id": "quran-recitations", "name": "Quran Recitations", "type": "dataset", "country": "INTL", "org": "MohamedRashad", "license": "unknown", "modality": "speech", "tasks": [ "asr", "quran" ], "links": { "hf": "https://huggingface.co/datasets/MohamedRashad/Quran-Recitations" }, "dialects": [ "classical" ], "size": "100K–1M rows", "year": 2025, "notes": "Quranic verses paired with audio recitations by multiple Qaris, 124,689 audio-text rows.", "metrics": { "downloads": 333, "likes": 63, "lastModified": "2025-03-30" } }, { "id": "covost2-en-ar", "name": "CoVoST2-EN-AR", "type": "dataset", "country": "INTL", "org": "ymoslem", "license": "cc-by-nc-4.0", "modality": "speech", "tasks": [ "asr", "tts", "translation", "speech" ], "links": { "hf": "https://huggingface.co/datasets/ymoslem/CoVoST2-EN-AR" }, "size": "100K–1M rows", "year": 2024, "notes": "CoVoST 2 is a large-scale multilingual speech translation corpus based on Common Voice, developed by FAIR.", "metrics": { "downloads": 332, "likes": 5, "lastModified": "2024-12-04" } }, { "id": "arabic-glm-ocr-v2", "name": "Arabic-GLM-OCR-v2", "type": "ocr", "country": "INTL", "org": "sherif1313", "license": "apache-2.0", "modality": "vision", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/sherif1313/Arabic-GLM-OCR-v2" }, "size": "1.1B", "year": 2026, "on_device": true, "dialects": [ "msa" ], "notes": "GLM-OCR fine-tuned for Arabic documents.", "base_model": [ "zai-org/glm-ocr" ], "metrics": { "downloads": 331, "likes": 17, "lastModified": "2026-05-27" } }, { "id": "cohere-transcribe-arabic-07-2026-dialectal", "name": "cohere transcribe arabic 07 2026 dialectal", "type": "asr", "country": "EG", "org": "oddadmix", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr", "diacritization" ], "links": { "hf": "https://huggingface.co/oddadmix/cohere-transcribe-arabic-07-2026-dialectal" }, "dialects": [ "mixed" ], "year": 2026, "notes": "Full fine-tune of CohereLabs/cohere-transcribe-arabic-07-2026 (2.07B, cohereasr Conformer encoder-decoder) for multi-dialect Arabic speech recognition.", "base_model": [ "coherelabs/cohere-transcribe-arabic-07-2026" ], "metrics": { "downloads": 330, "likes": 2, "lastModified": "2026-07-09" } }, { "id": "unimorph", "name": "UniMorph", "type": "dataset", "country": "INTL", "org": "Center for Language and Speech Processing Johns Hopkins University", "license": "cc-by-sa-3.0", "modality": "text", "tasks": [ "morphology" ], "links": { "github": "https://github.com/unimorph/ara", "hf": "https://huggingface.co/datasets/unimorph/universal_morphologies", "paper": "https://unimorph.github.io/doc/unimorph-schema.pdf" }, "dialects": [ "msa" ], "size": "140,003 tokens", "year": 2015, "notes": "167 languages have been annotated according to the UniMorph schema.", "metrics": { "downloads": 330, "likes": 20, "lastModified": "2023-06-08" } }, { "id": "tamazight-speech-to-arabic-text-tamazight-nlp", "name": "Tamazight Speech to Arabic Text", "type": "dataset", "country": "INTL", "org": "Tamazight-NLP", "license": "unknown", "modality": "speech", "tasks": [ "translation", "asr", "speech" ], "links": { "hf": "https://huggingface.co/datasets/Tamazight-NLP/Tamazight-Speech-to-Arabic-Text" }, "size": "10K–100K rows", "year": 2025, "notes": "This is the Tamazight-NLP organization-hosted version of the Tamazight-Arabic Speech Recognition Dataset.", "metrics": { "downloads": 324, "likes": 11, "lastModified": "2025-03-29" } }, { "id": "egyptian-asr-mgb3", "name": "Egyptian ASR MGB-3", "type": "dataset", "country": "EG", "org": "Mightystudent", "license": "unknown", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/datasets/MightyStudent/Egyptian-ASR-MGB-3" }, "size": "1K-10K rows", "year": 2023, "dialects": [ "egy" ], "notes": "Egyptian Arabic dialect ASR set derived from MGB-3 YouTube data.", "metrics": { "downloads": 322, "likes": 23, "lastModified": "2024-09-03" } }, { "id": "kde4", "name": "KDE4", "type": "dataset", "country": "INTL", "org": "Uppsala University", "license": "other", "modality": "text", "tasks": [ "translation" ], "links": { "hf": "https://huggingface.co/datasets/Helsinki-NLP/kde4", "paper": "http://www.lrec-conf.org/proceedings/lrec2012/pdf/463_Paper.pdf" }, "dialects": [ "msa" ], "size": "700,000 sentences", "year": 2012, "notes": "A parallel corpus of KDE4 localization files (v.2). 92 languages, 4,099 bitexts", "metrics": { "downloads": 322, "likes": 26, "lastModified": "2024-01-18" } }, { "id": "algerian-darija-text", "name": "Algerian Darija", "type": "dataset", "country": "DZ", "org": "ayoubkirouane", "license": "cc-by-4.0", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/datasets/ayoubkirouane/Algerian-Darija" }, "size": "100K-1M rows", "year": 2024, "dialects": [ "magh" ], "notes": "Algerian Darija text collected from HF datasets, scraping and social media (Algeria).", "metrics": { "downloads": 321, "likes": 17, "lastModified": "2026-04-18" } }, { "id": "quran-ocr", "name": "Quran OCR", "type": "benchmark", "country": "INTL", "org": "sherif1313", "license": "apache-2.0", "modality": "multimodal", "tasks": [ "ocr", "evaluation", "quran" ], "links": { "hf": "https://huggingface.co/datasets/sherif1313/Quran-OCR" }, "dialects": [ "classical" ], "size": "1K–10K rows", "year": 2026, "notes": "Any text recognition or Optical Character Recognition (OCR) system requires a dataset to learn how to recognize the text.", "metrics": { "downloads": 320, "likes": 0, "lastModified": "2026-02-18" } }, { "id": "ashaar-poetry", "name": "Ashaar", "type": "dataset", "country": "SA", "org": "ARBML", "license": "unknown", "modality": "text", "tasks": [ "pretraining" ], "links": { "hf": "https://huggingface.co/datasets/arbml/ashaar" }, "size": "254k poems", "year": 2022, "dialects": [ "classical" ], "notes": "Largest Arabic poetry dataset, 254k poems and 3.8M verses with meter and genre metadata.", "metrics": { "downloads": 317, "likes": 8, "lastModified": "2024-07-14" } }, { "id": "staro-ai-2-69-super", "name": "StarO-AI-2.69-Super", "type": "llm", "country": "INTL", "org": "C-a-Star-Technology-Official", "license": "creativeml-openrail-m", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/C-a-Star-Technology-Official/StarO-AI-2.69-Super" }, "year": 2026, "notes": "Created by Hawa-Al-Akram, Update By C.a.", "metrics": { "downloads": 317, "likes": 3, "lastModified": "2026-06-30" } }, { "id": "whisper-small-darija", "name": "whisper small darija", "type": "asr", "country": "MA", "org": "Youssef Chafiqui", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr", "evaluation" ], "links": { "hf": "https://huggingface.co/ychafiqui/whisper-small-darija" }, "dialects": [ "magh" ], "year": 2024, "notes": "This model is a fine-tuned version of openai/whisper-small on the Darija Speech to Text dataset.", "base_model": [ "openai/whisper-small" ], "metrics": { "downloads": 317, "likes": 3, "lastModified": "2024-09-27" } }, { "id": "nawah-50m-rag-chat", "name": "Nawah-50M RAG Chat", "type": "llm", "country": "EG", "org": "oddadmix", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/oddadmix/Nawah-50M-RAG-Chat-8K-GRPO" }, "size": "52M", "dialects": [ "msa" ], "notes": "Tiny 50M Arabic RAG chat model with 8K context, tuned with GRPO.", "base_model": [ "oddadmix/nawah-50m-rag-chat-8k" ], "metrics": { "downloads": 316, "likes": 0, "lastModified": "2026-08-20" } }, { "id": "arat5-base-title-generation", "name": "AraT5-base-title-generation", "type": "llm", "country": "INTL", "org": "UBC-NLP", "license": "unknown", "modality": "text", "tasks": [ "title-generation" ], "links": { "hf": "https://huggingface.co/UBC-NLP/AraT5-base-title-generation" }, "dialects": [ "msa" ], "year": 2022, "notes": "AraT5-base fine-tuned for Arabic title generation, from the AraT5 paper.", "metrics": { "downloads": 315, "likes": 13, "lastModified": "2022-05-26" } }, { "id": "mubsir-qwen-2b-vl", "name": "Mubsir Qwen 2B VL", "type": "ocr", "country": "INTL", "org": "Hatim2221", "license": "apache-2.0", "modality": "vision", "tasks": [ "chat", "asr", "ocr", "vision" ], "links": { "hf": "https://huggingface.co/Hatim2221/Mubsir-Qwen-2B-VL" }, "size": "2B", "on_device": true, "year": 2026, "notes": "This repository contains a vision-language model fine-tuned specifically for Arabic Handwritten Text Recognition (HTR).", "base_model": [ "qwen/qwen-vl" ], "metrics": { "downloads": 315, "likes": 5, "lastModified": "2026-08-24" } }, { "id": "arabic-audio-collection-algerian-kahwa-postcast", "name": "arabic audio collection algerian kahwa postcast", "type": "dataset", "country": "EG", "org": "oddadmix", "license": "other", "modality": "speech", "tasks": [ "tts", "asr", "diacritization", "sentiment", "speech" ], "links": { "hf": "https://huggingface.co/datasets/oddadmix/arabic-audio-collection-algerian-kahwa-postcast" }, "dialects": [ "magh" ], "size": "10K–100K rows", "year": 2026, "notes": "The Kahwa Postcast Arabic Speech Dataset is a large-scale, first-of-its-kind Arabic speech corpus containing approximately 110 hours of speech recordings.", "metrics": { "downloads": 313, "likes": 2, "lastModified": "2026-06-18" } }, { "id": "arabic-to-code-3m", "name": "Arabic-to-Code 8 Languages 3M", "type": "dataset", "country": "INTL", "org": "ISLAM-PO", "license": "cc-by-4.0", "modality": "text", "tasks": [ "instruction-tuning", "code" ], "links": { "hf": "https://huggingface.co/datasets/ISLAM-PO/arabic-to-code-8-langs-3m" }, "size": "3M", "year": 2026, "notes": "3M Arabic instruction examples for code generation across 8 languages.", "metrics": { "downloads": 313, "likes": 0, "lastModified": "2026-09-25" } }, { "id": "esprit-derja-qwen3-8b", "name": "ESPRIT-Derja-Qwen3-8B", "type": "llm", "country": "TN", "org": "ESPRIT Group", "license": "apache-2.0", "modality": "text", "tasks": [ "chat" ], "links": { "hf": "https://huggingface.co/ESPRIT-Group/ESPRIT-Derja-Qwen3-8B-v2" }, "base_model": [ "qwen/qwen3-8b" ], "size": "8.2B", "year": 2025, "dialects": [ "magh" ], "notes": "Qwen3-8B fine-tuned for Tunisian Derja by ESPRIT.", "metrics": { "downloads": 313, "likes": 2, "lastModified": "2026-06-08" } }, { "id": "ultradata-math-textbook-exercise-ar", "name": "ultradata math textbook exercise ar", "type": "dataset", "country": "INTL", "org": "SultanR", "license": "apache-2.0", "modality": "text", "tasks": [ "translation", "pretraining" ], "links": { "hf": "https://huggingface.co/datasets/SultanR/ultradata-math-textbook-exercise-ar" }, "size": "10M–100M rows", "year": 2026, "notes": "Arabic translation of the English portion of UltraData-Math, config UltraData-Math-L3-Textbook-Exercise-Synthetic.", "metrics": { "downloads": 312, "likes": 0, "lastModified": "2026-08-17" } }, { "id": "sard-extended", "name": "SARD Extended", "type": "dataset", "country": "SA", "org": "riotu-lab", "license": "cc-by-nc-nd-4.0", "modality": "vision", "tasks": [ "ocr" ], "links": { "hf": "https://huggingface.co/datasets/riotu-lab/SARD-Extended" }, "year": 2026, "notes": "SARD: large-scale synthetic Arabic OCR dataset rendered in many fonts (Amiri, Sakkal Majalla, Arial, Calibri and others).", "metrics": { "downloads": 311, "likes": 9, "lastModified": "2026-05-20" } }, { "id": "whisper-small-egyptian-codeswitch", "name": "Whisper Small Egyptian Code-Switch", "type": "asr", "country": "EG", "org": "Ibrahim Amin", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/IbrahimAmin/code-switched-egyptian-arabic-whisper-small" }, "size": "242M", "dialects": [ "egy", "mixed" ], "notes": "Whisper-small tuned for Egyptian Arabic with English code-switching.", "base_model": [ "openai/whisper-small" ], "metrics": { "downloads": 310, "likes": 5, "lastModified": "2025-05-17" } }, { "id": "ten2zero", "name": "Ten2Zero", "type": "dataset", "country": "SA", "org": "Umm Al-Qura University", "license": "cc-by-4.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/datasets/gfbati/Ten2Zero", "paper": "https://doi.org/10.21608/JESAUN.2023.231628.1254" }, "dialects": [ "mixed" ], "size": "935 tokens", "year": 2023, "notes": "Balanced spoken Arabic digit dataset (numbers 0-10) for teaching machine learning without coding, including wav files and generated tabular.", "metrics": { "downloads": 306, "likes": 1, "lastModified": "2023-10-20" } }, { "id": "dclm-edu-ar-500k", "name": "dclm edu ar 500k", "type": "dataset", "country": "INTL", "org": "SultanR", "license": "cc-by-4.0", "modality": "text", "tasks": [ "translation", "pretraining" ], "links": { "hf": "https://huggingface.co/datasets/SultanR/dclm-edu-ar-500k" }, "size": "100K–1M rows", "year": 2026, "notes": "This is an Arabic translation of the HuggingFaceTB/dclm-edu dataset.", "metrics": { "downloads": 305, "likes": 0, "lastModified": "2026-02-01" } }, { "id": "arabic-tweets", "name": "Arabic Tweets", "type": "dataset", "country": "INTL", "org": "Mohammad Albarham", "license": "cc-by-4.0", "modality": "text", "tasks": [ "nlp" ], "links": { "hf": "https://huggingface.co/datasets/pain/Arabic-Tweets" }, "notes": "41GB+ of Arabic tweets (~4B words)", "metrics": { "downloads": 304, "likes": 23, "lastModified": "2023-04-08" } }, { "id": "darija-speech-to-text", "name": "Darija Speech to Text", "type": "dataset", "country": "MA", "org": "adiren7", "license": "apache-2.0", "modality": "speech", "tasks": [ "asr" ], "links": { "hf": "https://huggingface.co/datasets/adiren7/darija_speech_to_text" }, "year": 2024, "dialects": [ "magh" ], "notes": "Moroccan Darija speech with transcripts for ASR training.", "metrics": { "downloads": 304, "likes": 17, "lastModified": "2024-08-11" } }, { "id": "ketaba-ocr-lora", "name": "Ketaba-OCR-LoRA", "type": "ocr", "country": "INTL", "org": "HassanB4", "license": "apache-2.0", "modality": "vision", "tasks": [ "ocr", "evaluation", "vision" ], "links": { "hf": "https://huggingface.co/HassanB4/Ketaba-OCR-LoRA" }, "year": 2026, "notes": "This repository contains the official models and results for Ketaba-OCR, our submission to the NakbaNLP 2026 Shared Task.", "base_model": [ "sherif1313/arabic-english-handwritten-ocr-v3" ], "metrics": { "downloads": 304, "likes": 5, "lastModified": "2026-03-05" } }, { "id": "mlma-hate-speech", "name": "MLMA hate speech", "type": "dataset", "country": "INTL", "org": "nedjmaou", "license": "mit", "modality": "text", "tasks": [ "nlp" ], "links": { "hf": "https://huggingface.co/datasets/nedjmaou/MLMA_hate_speech" }, "size": "10K–100K rows", "year": 2025, "notes": "Arabic text dataset (10K