LangSuitE: Planning, Controlling and Interacting with Large Language Models in Embodied Text Environments

TL;DR

LangSuitE introduces a versatile, simulation-free testbed for evaluating large language models in embodied tasks, emphasizing adaptability, internalized knowledge, and customizable interactions to advance embodied AI research.

Contribution

It presents LangSuitE, a novel testbed for embodied language tasks, and proposes EmMem, a chain-of-thought schema for summarizing embodied states, addressing key challenges in embodied planning.

Findings

LangSuitE effectively evaluates LLMs in diverse embodied tasks.

The EmMem schema improves the summarization of embodied states.

Benchmark results reveal significant challenges in embodied planning.

Abstract

Recent advances in Large Language Models (LLMs) have shown inspiring achievements in constructing autonomous agents that rely on language descriptions as inputs. However, it remains unclear how well LLMs can function as few-shot or zero-shot embodied agents in dynamic interactive environments. To address this gap, we introduce LangSuitE, a versatile and simulation-free testbed featuring 6 representative embodied tasks in textual embodied worlds. Compared with previous LLM-based testbeds, LangSuitE (i) offers adaptability to diverse environments without multiple simulation engines, (ii) evaluates agents' capacity to develop ``internalized world knowledge'' with embodied observations, and (iii) allows easy customization of communication and action strategies. To address the embodiment challenge, we devise a novel chain-of-thought (CoT) schema, EmMem, which summarizes embodied states w.r.t. history information. Comprehensive benchmark results illustrate challenges and insights of embodied planning. LangSuitE represents a significant step toward building embodied generalists in the context of language models.