Abstract:Recent advances in offline Reinforcement Learning (RL) have proven that effective policy learning can benefit from imposing conservative constraints on pre-collected datasets. However, such static datasets often exhibit distribution bias, resulting in limited generalizability. To address this limitation, a straightforward solution is data augmentation (DA), which leverages generative models to enrich data distribution. Despite the promising results, current DA techniques focus solely on reconstructing future trajectories from given states, while ignoring the exploration of history transitions that reach them. This single-direction paradigm inevitably hinders the discovery of diverse behavior patterns, especially those leading to critical states that may have yielded high-reward outcomes. In this work, we introduce Bidirectional Trajectory Diffusion (BiTrajDiff), a novel DA framework for offline RL that models both future and history trajectories from any intermediate states. Specifically, we decompose the trajectory generation task into two independent yet complementary diffusion processes: one generating forward trajectories to predict future dynamics, and the other generating backward trajectories to trace essential history this http URL can efficiently leverage critical states as anchors to expand into potentially valuable yet underexplored regions of the state space, thereby facilitating dataset diversity. Extensive experiments on the D4RL benchmark suite demonstrate that BiTrajDiff achieves superior performance compared to other advanced DA methods across various offline RL backbones.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2506.05762 [cs.LG] |
| (or arXiv:2506.05762v5 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2506.05762 arXiv-issued DOI via DataCite |
From: Yunpeng Qing [view email]
[v1]
Fri, 6 Jun 2025 05:41:33 UTC (471 KB)
[v2]
Fri, 29 Aug 2025 04:03:05 UTC (119 KB)
[v3]
Tue, 30 Dec 2025 09:13:03 UTC (846 KB)
[v4]
Mon, 12 Jan 2026 10:54:04 UTC (844 KB)
[v5]
Thu, 14 May 2026 17:01:38 UTC (1,467 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。