{"ID":2861092,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2510.03232","arxiv_id":"2510.03232","title":"LEAML: Label-Efficient Adaptation to Out-of-Distribution Visual Tasks for Multimodal Large Language Models","abstract":"Multimodal Large Language Models (MLLMs) have achieved strong performance on general visual benchmarks but struggle with out-of-distribution (OOD) tasks in specialized domains such as medical imaging, where labeled data is limited and expensive. We introduce LEAML, a label-efficient adaptation framework that leverages both scarce labeled VQA samples and abundant unlabeled images. Our approach generates domain-relevant pseudo question-answer pairs for unlabeled data using a QA generator regularized by caption distillation. Importantly, we selectively update only those neurons most relevant to question-answering, enabling the QA Generator to efficiently acquire domain-specific knowledge during distillation. Experiments on gastrointestinal endoscopy and sports VQA demonstrate that LEAML consistently outperforms standard fine-tuning under minimal supervision, highlighting the effectiveness of our proposed LEAML framework.","short_abstract":"Multimodal Large Language Models (MLLMs) have achieved strong performance on general visual benchmarks but struggle with out-of-distribution (OOD) tasks in specialized domains such as medical imaging, where labeled data is limited and expensive. We introduce LEAML, a label-efficient adaptation framework that leverages...","url_abs":"https://arxiv.org/abs/2510.03232","url_pdf":"https://arxiv.org/pdf/2510.03232v1","authors":"[\"Ci-Siang Lin\",\"Min-Hung Chen\",\"Yu-Yang Sheng\",\"Yu-Chiang Frank Wang\"]","published":"2025-10-03T17:59:56Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[\"Large Language Model\",\"Language Model\"]","has_code":false}