Integrating LLMs and Decision Transformers for Language Grounded Generative Quality-Diversity

TL;DR

This paper introduces a method combining Large Language Models and Decision Transformers to generate diverse, customizable trajectories in reinforcement learning environments, enabling natural language control and evaluation.

Contribution

It proposes a novel approach that integrates LLMs with decision transformers for language-grounded policy generation and introduces a new benchmark for evaluation.

Findings

Enables natural language specification of behaviors.

Allows customizable trajectory generation beyond discrete descriptors.

Provides an LLM-based evaluation method for generative agents.

Abstract

Quality-Diversity is a branch of stochastic optimization that is often applied to problems from the Reinforcement Learning and control domains in order to construct repertoires of well-performing policies/skills that exhibit diversity with respect to a behavior space. Such archives are usually composed of a finite number of reactive agents which are each associated to a unique behavior descriptor, and instantiating behavior descriptors outside of that coarsely discretized space is not straight-forward. While a few recent works suggest solutions to that issue, the trajectory that is generated is not easily customizable beyond the specification of a target behavior descriptor. We propose to jointly solve those problems in environments where semantic information about static scene elements is available by leveraging a Large Language Model to augment the repertoire with natural language descriptions of trajectories, and training a policy conditioned on those descriptions. Thus, our method allows a user to not only specify an arbitrary target behavior descriptor, but also provide the model with a high-level textual prompt to shape the generated trajectory. We also propose an LLM-based approach to evaluating the performance of such generative agents. Furthermore, we develop a benchmark based on simulated robot navigation in a 2d maze that we use for experimental validation.