Language models are robotic planners: reframing plans as goal refinement graphs

TL;DR

This paper introduces a method that uses goal refinement graphs generated by large language models to improve the correctness of robotic plans, leveraging goal modeling techniques from software engineering.

Contribution

It applies goal modeling techniques to LLM-generated plans, significantly enhancing their correctness and executability in robotic tasks.

Findings

Plans generated with goal refinement graphs are more correct.

Human judges find the programs more executable.

The approach outperforms previous methods in correctness.

Abstract

Successful application of large language models (LLMs) to robotic planning and execution may pave the way to automate numerous real-world tasks. Promising recent research has been conducted showing that the knowledge contained in LLMs can be utilized in making goal-driven decisions that are enactable in interactive, embodied environments. Nonetheless, there is a considerable drop in correctness of programs generated by LLMs. We apply goal modeling techniques from software engineering to large language models generating robotic plans. Specifically, the LLM is prompted to generate a step refinement graph for a task. The executability and correctness of the program converted from this refinement graph is then evaluated. The approach results in programs that are more correct as judged by humans in comparison to previous work.