Learning Where to Look: Observation Policy Optimization for Thinking with Images
cs.CV
Submitted: 2026-11-03
Updated: 2026-11-03
Code: https://github.com/Yann0122/Observation-Policy-Optimization
Terminology
Sources
- Qwen2.5-VL Technical Report
- Qwen3-VL Technical Report
- InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Models
- Kimi-VL Technical Report
- Seed1.5-VL Technical Report
- Qwen Look Again: Guiding Vision-Language Reasoning Models to Re-attention Visual Information
- DeepEyes: Incentivizing "Thinking with Images" via Reinforcement Learning
- GRIT: Teaching MLLMs to Think with Images
- Ground-R1: Incentivizing Grounded Visual Reasoning via Reinforcement Learning
- Thinking with Images for Multimodal Reasoning: Foundations, Methods, and Future Frontiers
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- Adaptive Chain-of-Focus Reasoning via Dynamic Visual Search and Zooming for Efficient VLMs
- ACTIVE-o3: Empowering MLLMs with Active Perception via Pure Reinforcement Learning
- OpenThinkIMG: Learning to Think with Images via Visual Tool Reinforcement Learning
- Thyme: Think Beyond Images
- DeepEyesV2: Toward Agentic Multimodal Model
- SenseNova-MARS: Empowering Multimodal Agentic Reasoning and Search via Reinforcement Learning
- Latent Visual Reasoning
- Zooming without Zooming: Region-to-Image Distillation for Fine-Grained Multimodal Perception
- Thinking with Programming Vision: Towards a Unified View for Thinking with Images
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs