AI Paper Atlas

Research map / Vision-language and multimodal / Visual question answering

Synthesizing High-Quality Visual Question Answering from Medical Documents with Generator-Verifier LMMs

Xiaoke Huang, Ningsen Wang, Hui Liu, Xianfeng Tang, Yuyin Zhou

ICLR 2026 · Vision-language and multimodal › Visual question answering · approach: Multimodal question answering datasets

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add rubric-guided verification to rubric-guided generation (ablation)MedVLSynther 3B · Six medical VQA benchmarks averageaccuracy (avg): 54.72 -> 55.85Improved
add rubric-guided verification to rubric-guided generation (ablation)MedVLSynther 7B · Six medical VQA benchmarks averageaccuracy (avg): 57.33 -> 57.56Improved
use rubric-guided context-aware generation instead of PMC-style text-only generation (ablation)MedVLSynther 7B · Six medical VQA benchmarks averageaccuracy (avg): 54.41 -> 57.33Improved
use rubric-guided generation instead of PMC-style text-only generation (ablation)MedVLSynther 3B · Six medical VQA benchmarks averageaccuracy (avg): 54.80 -> 54.72No clear effect
increase training data from 1K to 5K items (ablation)MedVLSynther 3B · Six medical VQA benchmarks averageaccuracy (avg): 52.64 -> 55.85Improved
increase training data beyond 5K to 10K and full (ablation)MedVLSynther 3B · Six medical VQA benchmarks averageaccuracy (avg): 55.85 -> 55.03 (10K), 55.17 (Full)No clear effect
increase training data from 5K to full 13K items (ablation)MedVLSynther 7B · Six medical VQA benchmarks averageaccuracy (avg): 57.56 -> 58.15Improved
use stronger verifier GLM-4.5V 108B as both generator and verifier (ablation)MedVLSynther 7B · Six medical VQA benchmarks averageaccuracy (avg): 57.56 -> 58.08Improved
use same GLM-4.5V model as generator and verifier instead of Qwen2.5-VL 72B verifier (ablation)MedVLSynther 3B · Six medical VQA benchmarks averageaccuracy (avg): 55.85 -> 54.59Hurt
use RL instead of SFT on MedSynVQA (ablation)Qwen2.5-VL 3B · Six medical VQA benchmarks averageaccuracy (avg): 49.11 -> 55.85Improved
use MedSynVQA instead of text-only m23k data for RL (ablation)Qwen2.5-VL 7B · Six medical VQA benchmarks averageaccuracy (avg): 54.88 -> 57.56Improved
use MedSynVQA instead of PMC-VQA data for RL (ablation)Qwen2.5-VL 3B · Six medical VQA benchmarks averageaccuracy (avg): 52.28 -> 55.85Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy