Awesome World Models / arXiv

IRL-VLA: Training an Vision-Language-Action Policy via Reward World Model

作者：作者信息待提取 · 单位：School of Communication and Information Engineering, Shanghai University, School of Mechanical Engineering, Shanghai Jiao Tong University, AIR, Tsinghua University, Beijing · 来源：World Models for Autonomous Driving · 发布日期：2025

自动驾驶具身智能视觉语言动作/策略学习语言先验视觉语言桥接
新标签打开 AlphaXiv新标签打开 Project新标签打开 Code
IRL-VLA: Training an Vision-Language-Action Policy via Reward World Model figure
AlphaXiv 中文概览（可滚动查看）