{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/catt-character-based-arabic-tashkeel","title":"CATT: Character-based Arabic Tashkeel Transformer","arxiv_id":"2407.03236","date":"2024-07-03","proceeding":null,"authors":["Faris Alasmary","Orjuwan Zaafarani","Ahmad Ghannam"],"abstract":"Tashkeel, or Arabic Text Diacritization (ATD), greatly enhances the comprehension of Arabic text by removing ambiguity and minimizing the risk of misinterpretations caused by its absence. It plays a crucial role in improving Arabic text processing, particularly in applications such as text-to-speech and machine translation. This paper introduces a new approach to training ATD models. First, we finetuned two transformers, encoder-only and encoder-decoder, that were initialized from a pretrained character-based BERT. Then, we applied the Noisy-Student approach to boost the performance of the best model. We evaluated our models alongside 11 commercial and open-source models using two manually labeled benchmark datasets: WikiNews and our CATT dataset. Our findings show that our top model surpasses all evaluated models by relative Diacritic Error Rates (DERs) of 30.83\\% and 35.21\\% on WikiNews and CATT, respectively, achieving state-of-the-art in ATD. In addition, we show that our model outperforms GPT-4-turbo on CATT dataset by a relative DER of 9.36\\%. We open-source our CATT models and benchmark dataset for the research community\\footnote{https://github.com/abjadai/catt}.","url_abs":"https://arxiv.org/abs/2407.03236v3","url_pdf":"https://arxiv.org/pdf/2407.03236v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"catt-character-based-arabic-tashkeel","repo_url":"https://github.com/abjadai/catt","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"arabic-text-diacritization","task_name":"Arabic Text Diacritization"},{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"machine-translation","task_name":"Machine Translation"},{"task_slug":"text-to-speech","task_name":"Text to Speech"},{"task_slug":"text-to-speech-1","task_name":"text-to-speech"}],"methods":[{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"attention-dropout","method_name":"Attention Dropout"},{"method_slug":"bert","method_name":"BERT"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"linear-warmup-with-linear-decay","method_name":"Linear Warmup With Linear Decay"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"weight-decay","method_name":"Weight Decay"},{"method_slug":"wordpiece","method_name":"WordPiece"}],"datasets_introduced":[{"slug":"catt-dataset","name":"CATT","full_name":"CATT Arabic Diacritization Benchmark Dataset"}],"methods_introduced":[],"results":[{"leaderboard":"/sota/arabic-text-diacritization-on-catt-dataset","task":"Arabic Text Diacritization","dataset":"CATT","model":"CATT ED","rank_in_archive_order":1,"of":12,"metrics":{"DER(%)":"8.624","WER (%)":"34.191"},"uses_additional_data":true},{"leaderboard":"/sota/arabic-text-diacritization-on-catt-dataset","task":"Arabic Text Diacritization","dataset":"CATT","model":"CATT EO","rank_in_archive_order":2,"of":12,"metrics":{"DER(%)":"8.762","WER (%)":"35.597"},"uses_additional_data":true},{"leaderboard":"/sota/arabic-text-diacritization-on-catt-dataset","task":"Arabic Text Diacritization","dataset":"CATT","model":"GPT-4","rank_in_archive_order":3,"of":12,"metrics":{"DER(%)":"9.515","WER (%)":"38.311"},"uses_additional_data":true},{"leaderboard":"/sota/arabic-text-diacritization-on-catt-dataset","task":"Arabic Text Diacritization","dataset":"CATT","model":"CBHG","rank_in_archive_order":4,"of":12,"metrics":{"DER(%)":"10.808","WER (%)":"42.680"},"uses_additional_data":false},{"leaderboard":"/sota/arabic-text-diacritization-on-catt-dataset","task":"Arabic Text Diacritization","dataset":"CATT","model":"Command R+","rank_in_archive_order":5,"of":12,"metrics":{"DER(%)":"13.169","WER (%)":"48.518"},"uses_additional_data":true},{"leaderboard":"/sota/arabic-text-diacritization-on-catt-dataset","task":"Arabic Text Diacritization","dataset":"CATT","model":"Shakkala","rank_in_archive_order":7,"of":12,"metrics":{"DER(%)":"13.494","WER (%)":"50.387"},"uses_additional_data":false},{"leaderboard":"/sota/arabic-text-diacritization-on-catt-dataset","task":"Arabic Text Diacritization","dataset":"CATT","model":"Sakhr","rank_in_archive_order":8,"of":12,"metrics":{"DER(%)":"13.841","WER (%)":"56.661"},"uses_additional_data":false},{"leaderboard":"/sota/arabic-text-diacritization-on-catt-dataset","task":"Arabic Text Diacritization","dataset":"CATT","model":"Alkhalil","rank_in_archive_order":9,"of":12,"metrics":{"DER(%)":"14.232 ","WER (%)":"53.413"},"uses_additional_data":false},{"leaderboard":"/sota/arabic-text-diacritization-on-catt-dataset","task":"Arabic Text Diacritization","dataset":"CATT","model":"Multilevel Diacritizer","rank_in_archive_order":10,"of":12,"metrics":{"DER(%)":"16.482","WER (%)":"60.844"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}