From Cooking Recipes to Robot Task Trees -- Improving Planning Correctness and Task Efficiency by Leveraging LLMs with a Knowledge Network

TL;DR

This paper presents a novel pipeline that leverages large language models and a knowledge network to generate accurate and efficient robot task plans for cooking, improving correctness and execution speed.

Contribution

It introduces a new method combining LLMs and task tree retrieval to enhance planning accuracy and efficiency in robotic cooking tasks.

Findings

Superior task planning accuracy over previous methods

Enhanced execution efficiency in cooking tasks

Effective mitigation of LLM output uncertainties

Abstract

Task planning for robotic cooking involves generating a sequence of actions for a robot to prepare a meal successfully. This paper introduces a novel task tree generation pipeline producing correct planning and efficient execution for cooking tasks. Our method first uses a large language model (LLM) to retrieve recipe instructions and then utilizes a fine-tuned GPT-3 to convert them into a task tree, capturing sequential and parallel dependencies among subtasks. The pipeline then mitigates the uncertainty and unreliable features of LLM outputs using task tree retrieval. We combine multiple LLM task tree outputs into a graph and perform a task tree retrieval to avoid questionable nodes and high-cost nodes to improve planning correctness and improve execution efficiency. Our evaluation results show its superior performance compared to previous works in task planning accuracy and efficiency.