AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Improving Reasoning in Vision-Language Models via Perception Verified Self-Training

Sourabh Sharma, Sonam Gupta, Sadbhawna Sadbhawna

ECCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Chain-of-thought multimodal reasoning · 1 citations (OpenAlex)

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 PerceptEval caption filtering to answer filtering (stage 1) (ablation)Stage-1 self-training · M3CoT Lang. Sci.accuracy (%): 49.76 -> 60.66Improved
add PerceptEval caption filtering to answer filtering (stage 1) (ablation)Stage-1 self-training · M3CoT Commonsenseaccuracy (%): 62.42 -> 69.45Improved
add two-stage curriculum with medium samples to dual-filter stage 1 (ablation)Ours · M3CoT Lang. Sci.accuracy (%): 60.66 -> 64.93Improved
add curriculum learning instead of mixing medium samples from the start (ablation)Ours · M3CoT Lang. Sci.accuracy (%): 57.34 -> 64.93Improved
add curriculum learning instead of mixing medium samples from the start (ablation)Ours · M3CoT Commonsenseaccuracy (%): 65.05 -> 72.74Improved
add FG-CLIP similarity to OCR-only PerceptEval (ablation)PerceptEval · M3CoT Commonsenseaccuracy (%): 60.44 -> 72.74Improved
add OCR similarity to FG-CLIP-only PerceptEval (ablation)PerceptEval · M3CoT Lang. Sci.accuracy (%): 56.87 -> 64.93Improved
add FG-CLIP similarity to OCR-only PerceptEval (ablation)PerceptEval · M3CoT Lang. Sci.accuracy (%): 61.13 -> 64.93Improved
add OCR similarity to FG-CLIP-only PerceptEval (ablation)PerceptEval · M3CoT Commonsenseaccuracy (%): 69.89 -> 72.74Improved
use perception-verified self-training instead of R3VLLaVA-v1.5-7B · M3CoT Lang. Sci.accuracy (%): 50.24 -> 64.93Improved
use perception-verified self-training instead of R3V on OCR hallucination splitLLaVA-v1.5-7B · HallusionBench OCR splitaccuracy (%): 55.24 -> 60.14Improved
use perception-verified self-training instead of R3V on out-of-domain dataLLaVA-v1.5-7B · MMMU Science valaccuracy (%): 33.78 -> 36.48Improved

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