Active Exploring like a Pigeon: Reinforcing Spatial Reasoning via Agentic Vision-Language Models 事件

Active Exploring like a Pigeon: Reinforcing Spatial Reasoning via Agentic Vision-Language Models arXiv:2606.02459v1 Announce Type: new Abstract: Enabling Vision-Language Models (VLMs) to perform spatial reasoning remains challenging. Existing approaches treat VLMs as passive observers, which is difficult for real-world applications. Moreover, reinforcement learning methods rely on sparse rewards, limiting their effectiveness for complex reasoning tasks. Inspired by pigeons' building and exploit