计划越多反而更糟：世界模型中的学习型规划器研究
When More Planning Hurts: Learned Planners for World Models
AI 导读
arXiv 论文《When More Planning Hurts》研究了世界模型中的学习型规划器。在 PushT 上，作者仅 1,505 参数的固定 action updater 在 16 个场景中解决 8 个，规划与编码总耗时 59.079s，而已发布的 LePlanner 解决 16 个任务、耗时仅 5.652s。
正文 · 原文
Abstract:On PushT, our fixed action updater (1,505 parameters) solves 8 of 16 scenes reserved from our configuration selection. The comparison planners perform as follows: Selected Adam (7 tasks solved, total planning and encoding time 218.440s), adapted iCEM (2 tasks solved, total planning and encoding time 49.309s), released LePlanner (16 tasks solved, total planning and encoding time 5.652s). The total planning and encoding time for our action updater is 59.079s. Note that our action updater does not satisfy the task-cost criterion we fixed before evaluation. We also conduct another matched experiment where we only replay the first ten commands in each context and then allow the next solve 4, 8 or 12 updates (subsequent solves all get 8 updates). Among all 16 contexts (8 scene clusters), the 3 budgets get 9, 8 and 7 successes, respectively. Interestingly, the extra updates did not rescue any failing task in this experiment. All the changes in success/failure happen in one scene cluster, where we observe that the extra updates decrease the costs of the selected latent world-model trajectories for our updater, while the tasks that used to be successes become failures. The code, weights, evaluators, execution traces and cost accounting accompany the paper.
| Comments: | 23 pages, 9 figures, 16 tables, 28 references |
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.37891 [cs.RO] |
| (or arXiv:2610.37891v1 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2610.37891 arXiv-issued DOI via DataCite |
Submission history
From: Josef Liyanjun Chen [view email]
[v1]
Tue, 27 Oct 2026 19:32:05 UTC (177 KB)
来源：arXiv cs.AI · arxiv.org