Online Learning of HTN Methods for integrated LLM-HTN Planning

TL;DR

This paper introduces an online learning approach for Hierarchical Task Network (HTN) methods integrated with LLM-based chatbots, enabling more efficient task decomposition and reducing reliance on ChatGPT during planning.

Contribution

It extends ChatHTN to learn generalized HTN methods from ChatGPT-generated decompositions, improving efficiency and applicability in integrated LLM-HTN planning.

Findings

Reduces ChatGPT calls during planning

Maintains or improves problem-solving success

Learns generalized methods from task decompositions

Abstract

We present online learning of Hierarchical Task Network (HTN) methods in the context of integrated HTN planning and LLM-based chatbots. Methods indicate when and how to decompose tasks into subtasks. Our method learner is built on top of the ChatHTN planner. ChatHTN queries ChatGPT to generate a decomposition of a task into primitive tasks when no applicable method for the task is available. In this work, we extend ChatHTN. Namely, when ChatGPT generates a task decomposition, ChatHTN learns from it, akin to memoization. However, unlike memoization, it learns a generalized method that applies not only to the specific instance encountered, but to other instances of the same task. We conduct experiments on two domains and demonstrate that our online learning procedure reduces the number of calls to ChatGPT while solving at least as many problems, and in some cases, even more.