arXiv：cs.LG（机器学习，全量分类）· Michael Hauri, Peter Buttaroni, Fabian A. Mikulasch, Friedemann Zenke·· 14 小时前AI 评分34
面向规划的保通勤时间世界模型 CTWM 提出
Learning Commute-Time-Preserving World Models for Planning
AI 导读
研究者提出保通勤时间世界模型（CTWMs），通过结合潜在位移预测器与对数行列式正则项防止表征坍缩，在可逆确定性动力学下可证明恢复正确缩放的图拉普拉斯表征。在多个复杂连续目标到达基准上，CTWM 匹配或超越任务无关基线 LeWM，且参数量仅为其一半。
正文
Abstract:World models allow agents to plan in latent space by choosing a sequence of actions that most reduces the distance to a given goal state. Thus, planning can benefit from latent representations whose distances mirror commute-times in the environment. The spectral embedding space of the graph Laplacian provides such a representation, if it obeys a specific eigenvalue-dependent scaling. Unfortunately, instantiating the graph Laplacian is intractable in large, continuous environments. Self-supervised learning offers a natural route to such commute-time-preserving embeddings at scale. However, here we show that existing methods, which commonly encourage isotropic representations to prevent representational collapse, tend to degrade the "correct" eigenvalue-dependent scaling, leading to an inaccurate representation of commute times. To address this problem, we introduce Commute-Time-Preserving World Models (CTWMs), combining a latent displacement predictor and a log-determinant regularizer that prevents collapse, which provably recover the correctly scaled Laplacian representation under reversible deterministic dynamics and at the predictor's fixed point. In numerical simulations, CTWM matches or outperforms LeWM, a task-agnostic baseline, on several complex, continuous goal-reaching benchmarks, while using half the parameters.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.01373 [cs.LG] |
| (or arXiv:2610.01373v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.01373 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Michael Hauri [view email]
[v1]
Thu, 1 Oct 2026 09:42:42 UTC (1,240 KB)
来源：arXiv：cs.LG（机器学习，全量分类） · arxiv.org