具身开发者社区

外观

论文arXiv 2610.0453925 天前

用于 Latent World Models 中可靠规划与自我改进的动作-结果对齐 ​

本文针对潜在世界模型预测误差低但规划不可靠的问题，提出ACA训练目标，并在多种环境与评估设置中验证其效果。

AI 整理
  • 硬件零部件
  • 数据与采集
  • 具身模型

原文链接：Action-Consequence Alignment for Reliable Planning and Self-Improving in Latent World Models（arXiv 2610.04539 · 发布日期：2026-10-03）

Action-Consequence Alignment for Reliable Planning and Self-Improving in Latent World Models

本文针对潜在世界模型预测误差低但规划不可靠的问题，提出ACA训练目标，并在多种环境与评估设置中验证其效果。

核心要点 ​

  • 指出相近替代动作可能预测误差更低，却产生更偏离记录目标的物理结果，导致预测与规划不匹配。
  • 提出 Action Consequence Alignment，在前向预测外惩罚局部搜索替代动作相对真实动作的预测误差优势。
  • ACA 训练不需额外模型组件或训练期间环境交互，并可用同一原则指导自我改进的数据收集。
  • 实验显示 ACA 提升规划性能、降低真实目标误差，ACA 引导的数据收集优于随机局部采样。

验证范围与局限：验证范围为摘要所述多种环境和评估设置，未说明具体环境、任务或是否包含真机。

英文摘要 (Abstract) ​

Latent world models learn to predict observed transitions, yet low prediction error alone does not guarantee reliable planning. Inspired by self tickling experiments in neuroscience showing that disrupting motor sensory correspondence increases prediction mismatch, we examine whether learned world models preserve an analogous action consequence correspondence.The results show nearby alternatives can receive lower prediction errors despite producing physical outcomes farther from the recorded target. With that future treated as a goal, this reveals a concrete prediction planning mismatch: the model assigns a lower cost to an action that achieves the target less accurately. To mitigate this gap, we introduce Action Consequence Alignment (ACA), a training objective that complements forward prediction by penalizing the prediction error advantage of locally searched alternatives over factual actions without additional model components or environment interactions during training. The same principle can also guide additional data collection for self improvement. We demonstrate that across diverse environments and evaluation settings, ACA improves planning performance and reduces real goal error, while ACA guided data collection outperforms random local sampling. These results support action consequence alignment as a practical principle for bridging predictive learning and reliable planning.

作者：Jinping Wang1, Zhiqiang Gao, Xiantong Zhen, Ling Shao

来源

本页事实对应的公开页面，以原文为准。

  1. arXiv 2610.04539：Action-Consequence Alignment for Reliable Planning and Self-Improving in Latent World Models访问 2026-10-21