arXiv：cs.AI· Wanjin Feng, Baobin Zhang, Ao Yu, Shibo Feng, Xi Wang, Xingyu Gao·· 6 小时前AI 评分39
PPWM：用于精准高效长时程规划的并行预测世界模型
Parallel Predictive World Models for Accurate and Efficient Long-Horizon Planning
AI 导读
研究者提出并行预测世界模型（PPWM），可并行预测有限时程轨迹，同时保留未来表征间的因果交互，从而去除自回归 rollout 中的解码状态反馈路径。在四个视觉控制任务上，PPWM 取得最低的长时程预测误差和最高的 CEM 模拟器成功率，CEM 规划速度平均较自回归 LeWM 基线提升超过 3 倍。
正文
Abstract:Long-horizon world-model planning typically relies on autoregressive rollouts, where predicted states are repeatedly fed back into the model. This preserves temporal structure but creates a horizon-length sequential path and exposes later predictions to recursive decoded-state feedback. We introduce Parallel Predictive World Models (PPWM), which predict a finite-horizon trajectory in parallel while retaining causal interaction among future representations. Each horizon is conditioned on its causal action prefix, and future representations interact before decoding, separating temporal causality from state-by-state output recursion. We formalize this distinction by viewing autoregressive rollout as a causal trajectory map and identifying the decoded-state feedback pathway removed by PPWM. Across four visual-control tasks, PPWM achieves the lowest long-horizon prediction error and the highest Cross-Entropy Method (CEM) simulator success among the evaluated predictive interfaces. Meanwhile, PPWM achieves more than a 3$\times$ average CEM planning speedup over the autoregressive LeWM baseline. These results suggest that accurate and efficient long-horizon world-model planning does not require state-by-state autoregression, but can instead be achieved through parallel causal trajectory prediction.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.08627 [cs.AI] |
| (or arXiv:2610.08627v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.08627 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Feng Wanjin [view email]
[v1]
Tue, 6 Oct 2026 16:26:34 UTC (217 KB)
来源：arXiv：cs.AI · arxiv.org