EmbodiExp：面向具身智能体的组合式经验迁移框架
EmbodiExp: Compositional Experience Transfer for Embodied Agents
AI 导读
研究者提出 EmbodiExp 框架，通过"结果锚定的子目标经验"为具身智能体构建可迁移经验，并在任务执行中在线组合。经验包含子目标、适用条件、可调整策略与可验证结果，失败轨迹仍可提供已验证的局部成功和需避免动作的指导。在多个具身基准上采用冻结经验库的 fold 式协议实验验证了其有效性。
正文 · 原文
Abstract:Experience can help embodied agents turn high-level plans into grounded actions, but transferring it across tasks remains challenging: previously effective strategies may be inapplicable to the current state or fail to produce their intended outcomes. We present EmbodiExp, a framework for constructing transferable experiences and composing them online during embodied task execution. Our key observation is that tasks often share subgoals whose completion can be assessed independently of overall task success. We therefore introduce an outcome-grounded subgoal experience that specifies a subgoal, applicability conditions, an adaptable policy, and a verifiable outcome. During execution, applicable experiences guide action generation, while verified outcomes can establish the conditions required by subsequent experiences, enabling online composition. Subgoal-level outcome attribution further enables failure utilization: failed trajectories can still contribute verified local successes and condition-specific guidance on actions to avoid. Experiments on multiple embodied benchmarks under a fold-based protocol with frozen experience banks demonstrate the effectiveness of EmbodiExp.
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.39180 [cs.RO] |
| (or arXiv:2610.39180v1 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2610.39180 arXiv-issued DOI via DataCite |
Submission history
From: Zhifeng Gu [view email]
[v1]
Thu, 29 Oct 2026 02:08:42 UTC (5,752 KB)
来源：arXiv cs.AI · arxiv.org