BiTrajDiff

BiTrajDiff is a bidirectional diffusion framework for offline reinforcement learning that models both future and history trajectories from intermediate states.

Recommended citation: Yunpeng Qing, Yixiao Chi, Shuo Chen, Shunyu Liu, Kelu Yao, Sixu Lin, Litao Liu, and Changqing Zou. BiTrajDiff: Bidirectional Trajectory Generation with Diffusion Models for Offline Reinforcement Learning. ICML 2026.
Download Paper