New decoding method boosts medical VQA for small vision-language models

Researchers have developed a new decoding method called Wasserstein Equilibrium Decoding, designed to improve the reliability of small vision-language models (2-8B) in medical visual question answering tasks. This approach extends game-theoretic decoding to handle open-ended medical VQA by using a semantically aware Wasserstein stopping criterion. The method achieves consistent improvements on datasets like VQA-RAD and PathVQA, enhancing accuracy and reducing inference iterations compared to traditional baselines. AI

IMPACT Enhances the reliability and efficiency of small vision-language models for specialized medical applications.

RANK_REASON The cluster contains an academic paper detailing a new method for improving AI model performance on a specific task. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• Gemma 3-4B
• Luca Hagen
• MedGemma 4B
• Medical Visual Question Answering
• PathVQA
• Qwen3-VL-2B
• VQA-RAD
• Wasserstein Equilibrium Decoding

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →