Research map / Vision-language and multimodal / Multimodal LLMs
See First, Reason Later: Mutual Information-Guided Reinforcement Learning for Vision-Language Models
Yin Zhang, Zonghan Wu, Jiaxuan Zhao, Junfeng Fang, Zengxiang Li, Kun Wang, Qingsong Wen, Yilei Shao
Read the paperCheck your idea against 137,217 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes a mutual-information-guided RLVR method for vision-language models to improve visual grounding before reasoning, based only on the truncated abstract.
Problem
VLMs suffer from visual perception errors and hallucinations that hurt answer accuracy in complex reasoning; existing RLVR methods waste sampling budget and provide weak visual grounding signals.
Approach
Uses mutual information between visual input and responses to guide RL training so the policy attends to visual evidence before reasoning ("see first, reason later"). Details beyond the truncated abstract are not stated.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add MI-guided selection and fork sampling (MIRL-8) vs DAPO-8 | DAPO · Six-benchmark average | accuracy: 66.27 -> 68.38 | Improved |
| add MI-guided selection (MIRL-12) vs DAPO-16 with fewer trajectories | DAPO · Six-benchmark average | accuracy: 69.76 -> 70.22 | Improved |
| MIRL-8 with 8 trajectories vs DAPO-12 with 12 | DAPO · Six-benchmark average | accuracy: 69.17 -> 68.38 | No clear effect |
| add top-MI selection instead of random selection (ablation) | MIRL-12 · Six-benchmark average | accuracy: 68.74 -> 70.22 | Improved |
| add top-MI selection instead of bottom-MI selection (ablation) | MIRL-12 · Six-benchmark average | accuracy: 68.02 -> 70.22 | Improved |
| add top-MI selection instead of random selection (ablation) | MIRL-12 · Math benchmarks | accuracy: 72.58 -> 73.35 | Improved |
| add top-MI selection instead of random selection (ablation) | MIRL-12 · CoT | accuracy: 70.92 -> 74.35 | Improved |
| add top-MI selection instead of random selection (ablation) | MIRL-12 · General VL | accuracy: 61.88 -> 63.46 | Improved |
| add MI reward on description tokens (lambda 0.1) to task-only reward (ablation) | MIRL-12 · Six-benchmark average | accuracy: 68.37 -> 70.22 | Improved |
| use decoupled lambda 0.1 instead of MI-only lambda 1.0 (ablation) | MIRL-12 · Six-benchmark average | accuracy: 68.81 -> 70.22 | Improved |
| add MI reward to task-only reward (ablation) | MIRL-12 · General VL | accuracy: 65.23 -> 63.46 | Hurt |
| MI-only reward vs task-only reward (ablation) | MIRL-12 · Six-benchmark average | accuracy: 68.37 -> 68.81 | No clear effect |
Experimental setup
Builds on: RLVR
Closest papers by meaning
- Improving Vision-language Models with Perception-centric Process Reward Models (CVPR 2026 · similarity 0.92)
- PDCR: Perception-Decomposed Confidence Reward for Vision-Language Reasoning (CVPR 2026 · similarity 0.92)
- Bad Seeing or Bad Thinking? Rewarding Perception for Multimodal Reasoning (ICML 2026 · similarity 0.90)
- Perception-R1: Advancing Multimodal Reasoning Capabilities of MLLMs via Visual Perception Reward (ICLR 2026 · similarity 0.90)
- ViCrit: A Verifiable Reinforcement Learning Proxy Task for Visual Perception in VLMs (NeurIPS 2025 · similarity 0.90)
- SSL4RL: Revisiting Self-supervised Learning as Intrinsic Reward for Visual-Language Reasoning (ICML 2026 · similarity 0.90)
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.
