InterPreT：基于语言反馈的交互式谓词学习用于可泛化任务规划
InterPreT: Interactive Predicate Learning from Language Feedback for Generalizable Task Planning
浏览论文内容
中文总结 AI 辅助
提出 InterPreT 框架，通过 LLM 在具身交互中从人类语言反馈学习符号谓词与算子并编译为 PDDL 领域，实现长时程机器人规划，在复杂新任务上泛化性强，成功率显著优于基线。
中文摘要 AI 辅助
学习抽象状态表示和知识对于长时程机器人规划至关重要。我们提出了 InterPreT，这是一个由 LLM 驱动的框架，使机器人能够在具身交互过程中从人类非专家的语言反馈中学习符号谓词。学习到的谓词提供了环境状态的关系抽象，促进了捕获动作前提条件和效果的符号算子的学习。通过将学习到的谓词和算子即时编译成 PDDL 领域，InterPreT 允许使用 PDDL 规划器对任意领域内目标进行有效规划。在模拟和真实世界的机器人操作领域中，我们证明了 InterPreT 能够可靠地揭示控制环境动力学的关键谓词和算子。尽管从简单的训练任务中学习，这些谓词和算子对具有显著更高复杂度的新任务展现出强大的泛化能力。在最具挑战性的泛化设置中，InterPreT 在模拟中达到 73% 的成功率，在真实世界中达到 40%，大幅优于基线方法。
英文摘要
Learning abstract state representations and knowledge is crucial for long-horizon robot planning. We present InterPreT, an LLM-powered framework for robots to learn symbolic predicates from language feedback of human non-experts during embodied interaction. The learned predicates provide relational abstractions of the environment state, facilitating the learning of symbolic operators that capture action preconditions and effects. By compiling the learned predicates and operators into a PDDL domain on-the-fly, InterPreT allows effective planning toward arbitrary in-domain goals using a PDDL planner. In both simulated and real-world robot manipulation domains, we demonstrate that InterPreT reliably uncovers the key predicates and operators governing the environment dynamics. Although learned from simple training tasks, these predicates and operators exhibit strong generalization to novel tasks with significantly higher complexity. In the most challenging generalization setting, InterPreT attains success rates of 73% in simulation and 40% in the real world, substantially outperforming baseline methods.
发表机构
- University of California, Los Angeles（加州大学洛杉矶分校）
- The University of Texas at Austin（德克萨斯大学奥斯汀分校）
机构由 AI 辅助整理，请以论文原文为准。
