--- title: "BiMediX2 : Bio-Medical EXpert LMM for Diverse Medical Modalities" description: "BiMediX2 is a medical assistant model that works with both text and images and converses in English… 1 reference, cited by 3 papers here." authors: ["Sahal Shaji Mullappilly", "Mohammed Irfan Kurpath", "Sara Pieri", "Saeed Yahya Alseiari", "Shanavas Cholakkal", "Khaled M Aldahmani", "Fahad Shahbaz Khan", "Rao Muhammad Anwer", "Salman Khan", "Timothy Baldwin", "Hisham Cholakkal"] year: 2025 venue: "Findings of the Association for Computational Linguistics: EMNLP 2025" source: "https://aclanthology.org/2025.findings-emnlp.756/" doi: "10.18653/v1/2025.findings-emnlp.756" licence: "CC BY 4.0 (ACL Anthology)" licence_url: "https://creativecommons.org/licenses/by/4.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/en/papers/mullappilly-2025-bimedix2-bio-medical-expert-lmm/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/mullappilly-2025-bimedix2-bio-medical-expert-lmm/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2025](https://md.research.elamur.ai/free/en/papers/2025/page/11/) / BiMediX2 : Bio-Medical EXpert LMM for Diverse Medical Modalities # BiMediX2 : Bio-Medical EXpert LMM for Diverse Medical Modalities **Authors:** [Sahal Shaji Mullappilly](https://md.research.elamur.ai/free/en/authors/sahal-shaji-mullappilly/), Mohammed Irfan Kurpath, Sara Pieri, Saeed Yahya Alseiari, Shanavas Cholakkal, Khaled M Aldahmani, [Fahad Shahbaz Khan](https://md.research.elamur.ai/free/en/authors/fahad-shahbaz-khan/), [Rao Muhammad Anwer](https://md.research.elamur.ai/free/en/authors/rao-muhammad-anwer/), [Salman Khan](https://md.research.elamur.ai/free/en/authors/salman-khan/), [Timothy Baldwin](https://md.research.elamur.ai/free/en/authors/timothy-baldwin/) and [Hisham Cholakkal](https://md.research.elamur.ai/free/en/authors/hisham-cholakkal/) · **Year:** 2025 · **Venue:** [Findings of the Association for Computational Linguistics: EMNLP 2025](https://md.research.elamur.ai/free/en/venues/findings/) ([Findings of the Association for Computational Linguistics 2025](https://md.research.elamur.ai/free/en/venues/findings/2025/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/) **Links:** [Publisher's page](https://aclanthology.org/2025.findings-emnlp.756/) · [PDF](https://aclanthology.org/2025.findings-emnlp.756.pdf) · [DOI 10.18653/v1/2025.findings-emnlp.756](https://doi.org/10.18653/v1/2025.findings-emnlp.756) · [OpenAlex](https://openalex.org/W4416034057) **Citations:** 2 citations (OpenAlex, 2026-10-14) · [3](https://md.research.elamur.ai/free/en/papers/mullappilly-2025-bimedix2-bio-medical-expert-lmm/cited-by/) in our library, of the 149,439 papers whose bibliographies we read ## Summary BiMediX2 is a medical assistant model that works with both text and images and converses in English and Arabic. It targets two gaps in earlier medical vision-language systems. Adding image support often weakened their text-only medical skills, and most covered only English or a narrow set of imaging types. The authors built BiMed-V, a set of roughly 1.6M bilingual instructions that mixes text-only and image-text samples. The Arabic part was produced by translating English data with GPT-4o, and bilingual clinicians checked a random sample. They then trained a LLaVA-style model on Llama 3.1 in two stages. First a projector learns to map medical images into the language model's input space, and then LoRA instruction tuning is applied. They also released BiMed-MBench, an expert-verified bilingual benchmark for medical multimodal models. The model leads open-source competitors on that benchmark, by about 9% in English and over 20% in Arabic. It also does well on medical VQA, report generation and summarization, and text QA. This shows that one checkpoint can cover many tasks and modalities without losing its text competence. ## Key points - **Data:** - BiMed-V contains about 1.6M instructions. Sources include PMC-OA, Rad-VQA, Path-VQA and SLAKE. - It adds 163k VQA items rebuilt from the LLaVA-Med 60K-IM set. - Over 10k LLaVA-Med captions were turned into dialogues with Llama 3.1 70B. - **Training:** Stage 1 trains only the projector on 467K medical image-caption pairs. Stage 2 applies LoRA tuning on BiMed-V. Released sizes are 4B, 8B and 70B. - **Benchmark:** BiMed-MBench has 386 expert-verified queries in English and Arabic. It covers CT, MRI, CXR, histology and gross pathology images. - **Text benchmarks:** The 70B model averages 84.6 on the clinical LLM benchmarks (MMLU medical subsets, MedMCQA, MedQA, USMLE, PubMedQA). Compared with GPT-4, it scores more than 8% higher on USMLE and about 9% higher on UPHILL factual accuracy. - **Translation baseline:** An English-only 8B model with Google, Bing or Alibaba translation scores 44.9–45.3 on Arabic BiMed-MBench. The natively bilingual model scores 50.47. ## Interesting ideas - The authors tested their English model behind commercial translation systems, which shows whether native bilingual training adds anything beyond a cheap cascade. - In the staged ablation, text-only scores stayed exactly the same after the projector-only stage, which confirms that this stage left the language model untouched. - Training data deliberately mixed text-only and image-text instructions so that the original text skills would not erode under the new modality. - The authors used a strong LLM to turn flat image captions into multi-turn dialogues, which converts cheap static annotations into richer interactive training data. - The Arabic data was machine-translated in bulk, and experts checked only a random subset, which gives quality assurance at a fraction of full review cost. ## For Mova How these ideas could serve Mova, our small open language model — its files, stages and tests: - **I1** (learn): Every new learned component could be benchmarked against a trivial cascade or pipeline baseline as a negative control, which proves the added machinery earns its cost. - **I2** (tree): When gparse gains a new input feature or adapter, metrics on old treebanks could be checked for bit-identical results while the old weights stay frozen, which catches leakage between stages. - **I3** (tree): When adapting the parser to a new UD dialect or domain, a share of the original treebank data could be kept in each batch to prevent forgetting. - **I4** (learn): An LLM teacher tier could expand fable or school-text sentences into question-answer exchanges, giving the answer readers silver training material. - **I5** (learn): Silver data agreed by the witness parsers could be audited the same way, by hand-checking a random sample to estimate error rates before training. *Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.* ## Institutions - [Mohamed bin Zayed University of Artificial Intelligence](https://md.research.elamur.ai/free/en/universities/mohamed-bin-zayed-university-of-artificial-intelligence/) — University, Abu Dhabi, United Arab Emirates · 124 works · 5,676 citations - [Linköping University](https://md.research.elamur.ai/free/en/universities/linkoping-university/) — University, Linköping, Sweden · 18 works · 2,319 citations Also on the byline: [Government Medical College](https://ror.org/04aznd361) · [Tawam Hospital](https://ror.org/007a5h107) · [Sheikh Shakhbout Medical City](https://ror.org/00gk5fa11). 📚 Shelves (6): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [👁️ Computer Vision and Pattern Recognition](https://md.research.elamur.ai/free/en/shelves/computer-science/computer-vision-and-pattern-recognition/) · 2,869 works · [🧬 Biochemistry, Genetics and Molecular Biology](https://md.research.elamur.ai/free/en/shelves/biochemistry-genetics-and-molecular-biology/) · 234 works · [🧬 Molecular Biology](https://md.research.elamur.ai/free/en/shelves/biochemistry-genetics-and-molecular-biology/molecular-biology/) · 231 works · [🩺 Medicine](https://md.research.elamur.ai/free/en/shelves/medicine/) · 241 works · [🩺 Health Informatics](https://md.research.elamur.ai/free/en/shelves/medicine/health-informatics/) · 184 works 🧭 Topics (2): [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) · 2,908 works · [🖼️ Vision & language](https://md.research.elamur.ai/free/en/topics/multimodal/) · 657 works ## Cite Cite as: Sahal Shaji Mullappilly, Mohammed Irfan Kurpath, Sara Pieri, Saeed Yahya Alseiari, Shanavas Cholakkal, Khaled M Aldahmani, Fahad Shahbaz Khan, Rao Muhammad Anwer, Salman Khan, Timothy Baldwin and Hisham Cholakkal. 2025. BiMediX2 : Bio-Medical EXpert LMM for Diverse Medical Modalities. Findings of the Association for Computational Linguistics: EMNLP 2025. The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @inproceedings{mullappilly2025bimedix2, title = {{BiMediX2 : Bio-Medical EXpert LMM for Diverse Medical Modalities}}, author = {Sahal Shaji Mullappilly, Mohammed Irfan Kurpath, Sara Pieri, Saeed Yahya Alseiari, Shanavas Cholakkal, Khaled M Aldahmani, Fahad Shahbaz Khan, Rao Muhammad Anwer, Salman Khan, Timothy Baldwin and Hisham Cholakkal}, booktitle = {Findings of the Association for Computational Linguistics: EMNLP 2025}, year = {2025}, doi = {10.18653/v1/2025.findings-emnlp.756}, url = {https://doi.org/10.18653/v1/2025.findings-emnlp.756} } ``` - [← previous Bilingual BSARD: Extending Statutory Article Retrieval to Dutch](https://md.research.elamur.ai/free/en/papers/lotfi-2025-bilingual-bsard-extending-statutory-article/) - [next → Conflicting Needles in a Haystack: How LLMs behave when faced with contradictory information](https://md.research.elamur.ai/free/en/papers/kurfali-2025-conflicting-needles-haystack-how-llms/) - [↑ back to papers of 2025](https://md.research.elamur.ai/free/en/papers/2025/page/11/) ## References 1 references, 1 of them with a card of ours. ↗ leads to the work. 1. Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. 2023. [Gpt-4 technical report](https://md.research.elamur.ai/free/en/papers/openai-2023-gpt-4-technical-report/). arXiv preprint arXiv:2303.08774. [↗](https://arxiv.org/abs/2303.08774) *Updated 31 October 2026*