Research map / Vision-language and multimodal / Multimodal LLMs
Vision-G1: Towards General Reasoning Vision-Language Models via Reinforcement Learning
Yuheng Zha, Kun Zhou, Yujia Wu, Yushu Wang, Jie Feng, Zhi Xu, Shibo Hao, Zhengzhong Liu, Eric P. Xing, Zhiting Hu
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.
Vision-G1 applies reinforcement learning from verifiable rewards to a vision-language model to improve general visual reasoning across diverse domains, addressing the scarcity of verifiable reward data beyond math and logic.
Problem
Current VLMs trained with RLVR cover a narrow range of reasoning tasks (mainly math and logic) because verifiable reward data is scarce in broader domains, so they generalize poorly.
Approach
Uses reinforcement learning from verifiable rewards to train a vision-language model for general reasoning across broader visual domains, targeting the lack of verifiable reward data outside math and logic. Abstract only; method details unavailable.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| RL training with Vision-G1 pipeline | Qwen2.5-VL-7B · MathVista | accuracy: 65.0 -> 76.1 | Improved |
| RL training with Vision-G1 pipeline | Qwen2.5-VL-7B · Muir-Bench | accuracy: 39.8 -> 61.5 | Improved |
| RL training with Vision-G1 pipeline | Qwen2.5-VL-7B · VQA-RAD | accuracy: 74.5 -> 72.1 | Hurt |
| RL training with Vision-G1 pipeline | Qwen2.5-VL-7B · MMMU-Val | accuracy: 58.6 -> 53.4 | Hurt |
| add multi-round RL training (ablation) | Vision-G1 · MathVista | accuracy: 74.9 -> 76.1 | Improved |
| add multi-round RL training (ablation) | Vision-G1 · ChartQA-Pro | accuracy: 48.4 -> 47.7 | Hurt |
| add influence-based filtering and difficulty selection (ablation) | Vision-G1 · MathVista | accuracy: 71.5 -> 76.1 | Improved |
| add influence-based filtering and difficulty selection (ablation) | Vision-G1 · Muir-Bench | accuracy: 59.4 -> 61.5 | Improved |
| add domain-specific datasets (ablation) | Vision-G1 · MathVista | accuracy: 76.3 -> 76.1 | Hurt |
| add domain-specific datasets (ablation) | Vision-G1 · VQA-RAD | accuracy: 68.1 -> 72.1 | Improved |
| RL training on math-focused data (Vision-R1-7B) | Qwen2.5-VL-7B · VQA-RAD | accuracy: 74.5 -> 64.9 | Hurt |
Limitations
- Full text unavailable; evaluation details not known from abstract
Closest papers by meaning
- Reason-RFT: Reinforcement Fine-Tuning for Visual Reasoning of Vision Language Models (NeurIPS 2025 · similarity 0.91)
- GFlowVLM: Enhancing Multi-step Reasoning in Vision-Language Models with Generative Flow Networks (CVPR 2025 · similarity 0.91)
- Synthesizing Multimodal Verifiable Game Data to Boost VLMs' General Reasoning (ICLR 2026 · similarity 0.91)
- Activating Visual Context and Commonsense Reasoning Through Masked Prediction in VLMs (AAAI 2026 · similarity 0.91)
- Fine-Tuning Large Vision-Language Models as Decision-Making Agents via Reinforcement Learning (NeurIPS 2024 · similarity 0.91)
- Unveiling the Compositional Ability Gap in Vision-Language Reasoning Model (NeurIPS 2025 · 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.
