analiza aiqbench
This research addresses a critical gap in clinical AI by quantifying how large vision-language models degrade when processing medical queries in non-English languages. The authors introduce a training-free inference-time method that leverages English model capabilities to steer representations, successfully mitigating performance drops across eight languages and diverse clinical scenarios. For developers of multilingual health systems, this offers a practical pathway to ensure equitable diagnostic support for linguistically diverse patient populations without requiring costly retraining.
Analyzing and Mitigating Cross-Lingual Degradation in Multilingual Medical VQA
arXiv:2608.22363v1 Announce Type: new Abstract: Medical visual question answering (VQA) is a crucial task in clinical AI, yet its evaluation has so far centered almost exclusively on English, limiting its relevance to linguistically diverse patients and clinicians. Recent multilingual medical VQA benchmarks show that large vision-language models (LVLMs) degrade in non-English languages, but lack a fine-grained analysis of how cross-lingual variation affects the distinct capabilities that medical VQA requires. To this end, we construct a multilingual medical VQA benchmark over eight languages, organized into four representative scenarios that isolate the core capabilities medical VQA requires. Evaluating five open- and closed-source LVLMs, we find that cross-lingual degradation is not uniform but highly scenario-dependent. We therefore propose MedVL-XLRepE, a training-free scenario-aware representation engineering method, leveraging LVLMs' superior English medical VQA capability to steer non-English representations toward their English counterparts at inference time. Across three LVLMs and eight languages, MedVL-XLRepE consistently mitigates cross-lingual degradation, with gains of up to 6.33\%.
Abstract:Medical visual question answering (VQA) is a crucial task in clinical AI, yet its evaluation has so far centered almost exclusively on English, limiting its relevance to linguistically diverse patients and clinicians. Recent multilingual medical VQA benchmarks show that large vision-language models (LVLMs) degrade in non-English languages, but lack a fine-grained analysis of how cross-lingual variation affects the distinct capabilities that medical VQA requires. To this end, we construct a multilingual medical VQA benchmark over eight languages, organized into four representative scenarios that isolate the core capabilities medical VQA requires. Evaluating five open- and closed-source LVLMs, we find that cross-lingual degradation is not uniform but highly scenario-dependent. We therefore propose MedVL-XLRepE, a training-free scenario-aware representation engineering method, leveraging LVLMs' superior English medical VQA capability to steer non-English representations toward their English counterparts at inference time. Across three LVLMs and eight languages, MedVL-XLRepE consistently mitigates cross-lingual degradation, with gains of up to 6.33\%.
Artificial Intelligence (cs.AI)
arXiv:2608.22363 [cs.AI]
(or arXiv:2608.22363v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2608.22363 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 246 cuvinte
Subiecte conexe
LLM Safety Alignment in Low-Resource Languages: A Systematic Literature Review
arXiv · 19 aug. 2026
Inference-Time Steering for Cross-Lingual Factual Consistency in LLMs
arXiv · 23 iul. 2026
Probing the Geometry of Color Representations in Multilingual Language Models
arXiv · 14 oct. 2026
Quantifying Cross-Lingual Transfer in Paralinguistic Speech Tasks
arXiv · 7 oct. 2026