外观
论文arXiv 2610.2161213 天前
潜在世界模型中的内在动力学学习
原文链接：Learning Intrinsic Dynamics in Latent World Models（arXiv 2610.21612 · 发布日期：2026-10-15）
Learning Intrinsic Dynamics in Latent World Models
针对视觉变化掩盖动作—结果关系，提出以动作可预测性约束潜在表征，并在Push-T多外观潜在规划中验证Concat IDM。
核心要点
- 视觉外观变化可能干扰世界模型表征，掩盖与控制过程相关的内在动力学。
- 动作可预测性由连续潜在状态重构动作，旨在保留不同动作及后果的区分。
- Concat IDM用离线轨迹动作标签更新共享编码器，不改变测试时规划器。
- 在Push-T 24种外观三次训练中，规划成功率由9.1%升至77.2%。
验证范围与局限：验证范围为Push-T 24种外观的潜在规划实验，摘要未说明仿真或真机。
英文摘要 (Abstract)
Learning intrinsic dynamics---the relationship between actions and their consequences---is essential for world models. However, visual variations unrelated to the underlying control process can obscure this relationship in learned representations. We investigate the connection between action predictability, defined as the ability to reconstruct an action from consecutive latent states, and the preservation of intrinsic dynamics. Our central hypothesis is that encouraging action predictability discourages representations from discarding the distinctions between different actions and their consequences. We provide two complementary forms of evidence for this connection. Theoretically, we derive a positive lower bound on squared inverse prediction error when the encoder merges distinguishable outcomes of different actions from the same state and appearance, and show that forward prediction can still be exact on those transitions. Empirically, controlled appearance-diversity comparisons, supervision interventions, and latent probes support the role of action prediction in preserving control-relevant transition information. We instantiate action predictability with Concat IDM, which predicts actions from concatenated latent states and updates the shared encoder using action labels already available in offline trajectories, without changing the test-time planner. Across three independent training runs on Push-T with 24 appearances, Concat IDM increases planning success from 9.1% to 77.2%; its paired improvement is 52.9 percentage points larger than that observed with a single appearance. Together, these findings support action predictability as an inductive bias toward intrinsic dynamics and demonstrate the practical benefits of Concat IDM for latent planning under visual diversity.
作者：Jiamin Shi, Yuanhong Zhang, Zhiyang Liu, Jiawei Sun, Jingwen Fu
来源
本页事实对应的公开页面，以原文为准。