Branch-aware Graph Planning（BGP）如何突破离线目标强化学习中的路径幻象
Escaping Path Mirages in Offline Goal-Conditioned Reinforcement Learning
AI 导读
研究者提出 Branch-aware Graph Planning（BGP）方法，解决离线目标条件强化学习中的路径幻象问题，即智能体在随机环境下过度依赖数据集中虚假成功的轨迹。BGP 将高变异性点识别为图节点并设计目标条件边与分段学习，规避不可靠分支。在 OGBench 多项任务上，BGP 大幅超越此前 SOTA 方法，代码已开源，论文被 NeurIPS 2026 接收。
正文 · 原文
Abstract:Offline goal-conditioned reinforcement learning remains challenging in stochastic long-horizon settings, where compounding value estimation errors hinder reliable goal reaching. While prior methods have sought to address this challenge through various approaches, a key challenge remains path mirage, where agents overcommit to spuriously successful trajectories in offline datasets induced by stochasticity, often leading to failure on difficult tasks. To address this issue, we propose Branch-aware Graph Planning (BGP), which captures stochastic branching structures by identifying high-variability points as graph nodes and designing goal-conditioned edges with segmented learning to avoid unreliable branches and reduce unnecessary stochasticity along paths. As a result, BGP enables reliable goal reaching even in highly stochastic long-horizon environments, and experiments on diverse OGBench tasks show that it substantially outperforms prior state-of-the-art offline goal-conditioned RL methods. Our implementation is available at this https URL.
| Comments: | 10 pages (main), 34 pages in total. Accepted to NeurIPS 2026 |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.40130 [cs.LG] |
| (or arXiv:2610.40130v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.40130 arXiv-issued DOI via DataCite |
Submission history
From: Seungyul Han [view email]
[v1]
Thu, 29 Oct 2026 14:21:50 UTC (9,903 KB)
来源：arXiv cs.AI · arxiv.org