Agent World Model: Infinity Synthetic Environments for Agentic Reinforcement Learning
Published in arXiv preprint, 2026
Zhaoyang Wang, Canwen Xu, Boyi Liu, Yite Wang, Siwei Han, Zhewei Yao, Huaxiu Yao, and Yuxiong He.
This work proposes Agent World Model, which creates effectively unbounded synthetic environments for agentic reinforcement learning.