arXiv 2410.121242025-08-31cs.ROcs.AI

从10个演示中学习：基于定向可供性框架的可泛化且样本高效的策略学习

Learning from 10 Demos: Generalisable and Sample-Efficient Policy Learning with Oriented Affordance Frames

QUT Centre for Robotics, Queensland University of Technology（昆士兰科技大学机器人中心，昆士兰理工大学）；University of Adelaide（阿德莱德大学）

Krishan Rana, Jad Abou-Chakra, Sourav Garg, Robert Lee, Ian Reid, Niko Suenderhauf

备注：: Accepted at the Conference on Robot Learning (CoRL), 2025. Videos can be found on our project website: https://affordance-policy.github.io

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

该研究提出定向可供性框架这一结构化表示，仅需10个演示即可高效学习机器人策略，结合自进度预测实现子策略无缝切换，达成长时序多物体任务的组合泛化，在小样本下仍对未知物体属性与布局有高成功率。

摘要：

模仿学习已让机器人具备展现高灵巧行为的潜力，但受限于样本效率低下与泛化能力不足，在长时序、多物体任务中仍表现不佳。现有方法需要大量演示来覆盖可能的任务变化，成本高昂，在实际部署中往往不具备可行性。针对这一挑战，我们提出了定向可供性框架（oriented affordance frames），这是一种状态与动作空间的结构化表示，能够提升空间泛化与类别内泛化能力，仅需10个演示即可高效学习策略。更重要的是，我们证明该抽象表示可让独立训练的子策略实现组合泛化，以解决长时序、多物体任务。为实现子策略间的无缝切换，我们提出了自进度预测的概念，该概念直接由训练演示的时长推导得出。我们在三项真实世界任务上验证了方法的有效性，每项任务都需要多步骤、多物体交互。尽管数据集规模很小，我们的策略仍能稳健泛化到未见过的物体外观、几何形状与空间布局，无需依赖海量训练数据即可实现高成功率。视频演示可在我们的项目页面查看：https://affordance-policy.github.io/。

英文摘要：

Imitation learning has unlocked the potential for robots to exhibit highly dexterous behaviours. However, it still struggles with long-horizon, multi-object tasks due to poor sample efficiency and limited generalisation. Existing methods require a substantial number of demonstrations to cover possible task variations, making them costly and often impractical for real-world deployment. We address this challenge by introducing oriented affordance frames, a structured representation for state and action spaces that improves spatial and intra-category generalisation and enables policies to be learned efficiently from only 10 demonstrations. More importantly, we show how this abstraction allows for compositional generalisation of independently trained sub-policies to solve long-horizon, multi-object tasks. To seamlessly transition between sub-policies, we introduce the notion of self-progress prediction, which we directly derive from the duration of the training demonstrations. We validate our method across three real-world tasks, each requiring multi-step, multi-object interactions. Despite the small dataset, our policies generalise robustly to unseen object appearances, geometries, and spatial arrangements, achieving high success rates without reliance on exhaustive training data. Video demonstration can be found on our project page: https://affordance-policy.github.io/.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。