{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,18]],"date-time":"2026-06-18T22:02:18Z","timestamp":1781820138144,"version":"3.54.5"},"reference-count":33,"publisher":"MDPI AG","issue":"4","license":[{"start":{"date-parts":[[2026,4,21]],"date-time":"2026-04-21T00:00:00Z","timestamp":1776729600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["MAKE"],"abstract":"<jats:p>Dialectal variation presents a major challenge for deploying medical language models in real-world healthcare settings, where patient\u2013clinician communication often occurs in regional vernaculars rather than standardized language forms. This challenge is particularly pronounced in the Arabic-speaking world, where clinical interactions frequently take place in diverse dialects that differ substantially from Modern Standard Arabic. Fine-tuning and maintaining separate models for each dialect is computationally inefficient and difficult to scale, motivating more integrated approaches. In this work, we present MENARA, an Arabic medical language model constructed by merging Egyptian Arabic, Moroccan Darija, and medical-domain specialists through model merging. We extend prior feasibility findings through comprehensive evaluation of cross-dialect performance, medical safety, and cross-lingual knowledge retention. Specifically, we introduce a fine-grained dialect composition analysis to quantify lexical purity and structured code-switching behavior, benchmark against state-of-the-art Arabic LLMs, conduct subject-matter-expert assessment of both dialectal fidelity and medical appropriateness. The results show that model merging preserves core medical competence while enabling robust dialectal adaptation, achieving strong cross-dialect fidelity while substantially reducing storage and deployment overhead compared to maintaining separate models. These findings establish model merging as a potentially practical and resource-efficient paradigm for dialect-aware medical NLP in linguistically fragmented healthcare environments.<\/jats:p>","DOI":"10.3390\/make8040110","type":"journal-article","created":{"date-parts":[[2026,4,21]],"date-time":"2026-04-21T07:57:21Z","timestamp":1776758241000},"page":"110","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":1,"title":["MENARA: Medical Natural Arabic Response Assistant"],"prefix":"10.3390","volume":"8","author":[{"ORCID":"https:\/\/orcid.org\/0009-0002-9217-4173","authenticated-orcid":false,"given":"Ahmed","family":"Ibrahim","sequence":"first","affiliation":[{"name":"AI Innovation Lab, Weill Cornell Medicine\u2014Qatar, Doha P.O. Box 24144, Qatar"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2967-3033","authenticated-orcid":false,"given":"Abdullah","family":"Hosseini","sequence":"additional","affiliation":[{"name":"AI Innovation Lab, Weill Cornell Medicine\u2014Qatar, Doha P.O. Box 24144, Qatar"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0009-8251-9351","authenticated-orcid":false,"given":"Hoda","family":"Helmy","sequence":"additional","affiliation":[{"name":"AI Innovation Lab, Weill Cornell Medicine\u2014Qatar, Doha P.O. Box 24144, Qatar"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0815-7902","authenticated-orcid":false,"given":"Maryam","family":"Arabi","sequence":"additional","affiliation":[{"name":"AI Innovation Lab, Weill Cornell Medicine\u2014Qatar, Doha P.O. Box 24144, Qatar"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Aya","family":"AlShareef","sequence":"additional","affiliation":[{"name":"AI Innovation Lab, Weill Cornell Medicine\u2014Qatar, Doha P.O. Box 24144, Qatar"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2135-943X","authenticated-orcid":false,"given":"Wafa","family":"Lakhdhar","sequence":"additional","affiliation":[{"name":"AI Innovation Lab, Weill Cornell Medicine\u2014Qatar, Doha P.O. Box 24144, Qatar"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4145-5509","authenticated-orcid":false,"given":"Ahmed","family":"Serag","sequence":"additional","affiliation":[{"name":"AI Innovation Lab, Weill Cornell Medicine\u2014Qatar, Doha P.O. Box 24144, Qatar"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,4,21]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","unstructured":"Alasmari, A. (2025). A Scoping Review of Arabic Natural Language Processing for Mental Health. Healthcare, 13.","DOI":"10.3390\/healthcare13090963"},{"key":"ref_2","unstructured":"Muresan, S., Nakov, P., and Villavicencio, A. (2022). Morphosyntactic Tagging with Pre-trained Language Models for Arabic and its Dialects. Findings of the Association for Computational Linguistics: ACL 2022, Association for Computational Linguistics."},{"key":"ref_3","first-page":"104","article-title":"The Mutual Intelligibility of Arabic Dialects: Implications for the Language Classroom","volume":"8","author":"Trentman","year":"2020","journal-title":"Crit. Multiling. Stud."},{"key":"ref_4","doi-asserted-by":"crossref","unstructured":"Habash, N., Vogel, S., and Darwish, K. (2015). Natural Language Processing for Dialectical Arabic: A Survey. Proceedings of the Second Workshop on Arabic Natural Language Processing, Association for Computational Linguistics.","DOI":"10.18653\/v1\/W15-32"},{"key":"ref_5","doi-asserted-by":"crossref","unstructured":"Wu, X.K., Chen, M., Li, W., Wang, R., Lu, L., Liu, J., Hwang, K., Hao, Y., Pan, Y., and Meng, Q. (2025). Llm fine-tuning: Concepts, opportunities, and challenges. Big Data Cogn. Comput., 9.","DOI":"10.3390\/bdcc9040087"},{"key":"ref_6","first-page":"100658","article-title":"D3: A Small Language Model for Drug-Drug Interaction prediction and comparison with Large Language Models","volume":"20","author":"Ibrahim","year":"2025","journal-title":"Mach. Learn. Appl."},{"key":"ref_7","doi-asserted-by":"crossref","unstructured":"Ibrahim, A., Khalili, A., Arabi, M., Sattar, A., Hosseini, A., and Serag, A. (2025). MERA: Medical Electronic Records Assistant. Mach. Learn. Knowl. Extr., 7.","DOI":"10.3390\/make7030073"},{"key":"ref_8","first-page":"3","article-title":"Lora: Low-rank adaptation of large language models","volume":"1","author":"Hu","year":"2022","journal-title":"ICLR"},{"key":"ref_9","first-page":"1877","article-title":"Language models are few-shot learners","volume":"33","author":"Brown","year":"2020","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"ref_10","doi-asserted-by":"crossref","unstructured":"Brunet, G., Chechik, M., Easterbrook, S., Nejati, S., Niu, N., and Sabetzadeh, M. (2006). A manifesto for model merging. Proceedings of the 2006 International Workshop on Global Integrated Model Management, Shanghai, China, 22 May 2006, Association for Computing Machinery.","DOI":"10.1145\/1138304.1138307"},{"key":"ref_11","doi-asserted-by":"crossref","unstructured":"Xu, Z., Yuan, K., Wang, H., Wang, Y., Song, M., and Song, J. (2024). Training-free pretrained model merging. Proceedings of the 2024 IEEE\/CVF Conference on Computer Vision and Pattern Recognition, Seattle, WA, USA, 16\u201322 June 2024, Computer Vision Foundation.","DOI":"10.1109\/CVPR52733.2024.00565"},{"key":"ref_12","unstructured":"Darwish, K., Ali, A., Abu Farha, I., Touileb, S., Zitouni, I., Abdelali, A., Al-Ghamdi, S., Alkhereyf, S., Zaghouani, W., and Khalifa, S. (2025). Bridging Dialectal Gaps in Arabic Medical LLMs through Model Merging. Proceedings of the Third Arabic Natural Language Processing Conference, Association for Computational Linguistics."},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"2","DOI":"10.1016\/j.procs.2018.10.456","article-title":"A lexical distance study of Arabic dialects","volume":"142","author":"Kwaik","year":"2018","journal-title":"Procedia Comput. Sci."},{"key":"ref_14","doi-asserted-by":"crossref","unstructured":"Al-Wer, E., and de Jong, R. (2017). Dialects of Arabic. The Handbook of Dialectology, John Wiley & Sons, Inc.","DOI":"10.1002\/9781118827628.ch32"},{"key":"ref_15","unstructured":"Salameh, M., Bouamor, H., and Habash, N. (2018). Fine-grained Arabic dialect identification. Proceedings of the 27th International Conference on Computational Linguistics, Santa Fe, NW, USA, 20\u201326 August 2018, Association for Computational Linguistics."},{"key":"ref_16","unstructured":"Habash, N., Bouamor, H., Hajj, H., Magdy, W., Zaghouani, W., Bougares, F., Tomeh, N., Abu Farha, I., and Touileb, S. (2021). QADI: Arabic Dialect Identification in the Wild. Proceedings of the Sixth Arabic Natural Language Processing Workshop, Association for Computational Linguistics."},{"key":"ref_17","unstructured":"Darwish, K., Ali, A., Abu Farha, I., Touileb, S., Zitouni, I., Abdelali, A., Al-Ghamdi, S., Alkhereyf, S., Zaghouani, W., and Khalifa, S. (2025). AraHealthQA 2025: The First Shared Task on Arabic Health Question Answering. Proceedings of the Third Arabic Natural Language Processing Conference: Shared Tasks, Association for Computational Linguistics."},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"110855","DOI":"10.1016\/j.dib.2024.110855","article-title":"AHD: Arabic healthcare dataset","volume":"56","author":"Gawbah","year":"2024","journal-title":"Data Brief"},{"key":"ref_19","unstructured":"Habash, N., Bouamor, H., Hajj, H., Magdy, W., Zaghouani, W., Bougares, F., Tomeh, N., Abu Farha, I., and Touileb, S. (2021). ArCOV-19: The First Arabic COVID-19 Twitter Dataset with Propagation Networks. Proceedings of the Sixth Arabic Natural Language Processing Workshop, Association for Computational Linguistics."},{"key":"ref_20","doi-asserted-by":"crossref","unstructured":"Mohammad, R., Alkhnbashi, O.S., and Hammoudeh, M. (2024). Optimizing large language models for arabic healthcare communication: A focus on patient-centered NLP applications. Big Data Cogn. Comput., 8.","DOI":"10.3390\/bdcc8110157"},{"key":"ref_21","unstructured":"Yang, E., Shen, L., Guo, G., Wang, X., Cao, X., Zhang, J., and Tao, D. (2024). Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunities. arXiv."},{"key":"ref_22","doi-asserted-by":"crossref","first-page":"17703","DOI":"10.52202\/068431-1287","article-title":"Merging models with fisher-weighted averaging","volume":"35","author":"Matena","year":"2022","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"ref_23","unstructured":"Kodali, P., Shivkumar, V., Joshi, S., Choudhary, M., Kumaraguru, P., and Shrivastava, M. (2025). Adapting Multilingual Models to Code-Mixed Tasks via Model Merging. arXiv."},{"key":"ref_24","unstructured":"Wang, Y., Gu, Y., Zhang, Y., Zhou, Q., Yan, Z., Xie, C., Wang, X., Yuan, J., and Yang, H. (2025). Model Merging Scaling Laws in Large Language Models. arXiv."},{"key":"ref_25","unstructured":"Adelani, D.I., Arnett, C., Ataman, D., Chang, T.A., Gonen, H., Raja, R., Schmidt, F., Stap, D., and Wang, J. (2025). The Unreasonable Effectiveness of Model Merging for Cross-Lingual Transfer in LLMs. Proceedings of the 5th Workshop on Multilingual Representation Learning (MRL 2025), Association for Computational Linguistics."},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"132498","DOI":"10.1016\/j.neucom.2025.132498","article-title":"A differentiable and uncertainty-aware mutual information regularizer for bias mitigation","volume":"669","author":"Incremona","year":"2026","journal-title":"Neurocomputing"},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"1148","DOI":"10.1109\/TPAMI.2024.3487254","article-title":"FLAC: Fairness-Aware Representation Learning by Suppressing Attribute-Class Associations","volume":"47","author":"Sarridis","year":"2025","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref_28","first-page":"7093","article-title":"Ties-merging: Resolving interference when merging models","volume":"36","author":"Yadav","year":"2023","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"ref_29","unstructured":"Team, G., Mesnard, T., Hardin, C., Dadashi, R., Bhupatiraju, S., Pathak, S., Sifre, L., Rivi\u00e8re, M., Kale, M.S., and Love, J. (2024). Gemma: Open Models Based on Gemini Research and Technology. arXiv."},{"key":"ref_30","unstructured":"OpenMeditron (2026, April 13). Meditron3-Gemma2-2B. Hugging Face Model Repository. Available online: https:\/\/huggingface.co\/OpenMeditron\/Meditron3-Gemma2-2B."},{"key":"ref_31","unstructured":"MBZUAI-Paris (2026, April 13). Egyptian-SFT-Mixture Dataset. Available online: https:\/\/huggingface.co\/datasets\/MBZUAI-Paris\/Egyptian-SFT-Mixture."},{"key":"ref_32","doi-asserted-by":"crossref","unstructured":"Hettiarachchi, H., Ranasinghe, T., Rayson, P., Mitkov, R., Gaber, M., Premasiri, D., Tan, F.A., and Uyangodage, L. (2025). Atlas-Chat: Adapting Large Language Models for Low-Resource Moroccan Arabic Dialect. Proceedings of the First Workshop on Language Models for Low-Resource Languages, Association for Computational Linguistics.","DOI":"10.18653\/v1\/2026.loreslm-1.56"},{"key":"ref_33","unstructured":"Dernoncourt, F., Preo\u0163iuc-Pietro, D., and Shimorina, A. (2024). Arcee\u2019s MergeKit: A Toolkit for Merging Large Language Models. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing: Industry Track, Association for Computational Linguistics."}],"container-title":["Machine Learning and Knowledge Extraction"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2504-4990\/8\/4\/110\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,21]],"date-time":"2026-04-21T08:38:03Z","timestamp":1776760683000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2504-4990\/8\/4\/110"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,4,21]]},"references-count":33,"journal-issue":{"issue":"4","published-online":{"date-parts":[[2026,4]]}},"alternative-id":["make8040110"],"URL":"https:\/\/doi.org\/10.3390\/make8040110","relation":{},"ISSN":["2504-4990"],"issn-type":[{"value":"2504-4990","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,4,21]]}}}