Reinforcement Fine-Tuning of Flow-Matching Policies for Vision-Language-Action Models

作者：Mingyang Lyua, Yinqian Suna, Erliang Lina, Huangrui Lib, Ruolin Chena, Feifei Zhaoa, Yi Zenga · 单位：Automation, Chinese Academy of Sciences, Beijing, China, b University, of Chinese Academy of Sciences (UCAS), Beijing, China, Laboratory of Brain Cognition and Brain-inspired Intelligence Technology · 会议/期刊：arXiv · 日期：2025-10-11 · 来源：Low-Level Learning-Based Action Modelling / Input Modelling / 2D Vision Language Action Models with RL