Saved in:
Bibliographic Details
Main Authors: Lică, Mircea, Shirekar, Ojas, Colle, Baptiste, Raman, Chirag
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2411.12977
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908713546153984
author Lică, Mircea
Shirekar, Ojas
Colle, Baptiste
Raman, Chirag
author_facet Lică, Mircea
Shirekar, Ojas
Colle, Baptiste
Raman, Chirag
contents Embodied agents powered by large language models (LLMs), such as Voyager, promise open-ended competence in worlds such as Minecraft. However, when powered by open-weight LLMs they still falter on elementary tasks after domain-specific fine-tuning. We propose MindForge, a generative-agent framework for cultural lifelong learning through explicit perspective taking. We introduce three key innovations: (1) a structured theory of mind representation linking percepts, beliefs, desires, and actions; (2) natural inter-agent communication; and (3) a multi-component memory system. Following the cultural learning framework, we test MindForge in both instructive and collaborative settings within Minecraft. In an instructive setting with GPT-4, MindForge agents powered by open-weight LLMs significantly outperform their Voyager counterparts in basic tasks yielding $3\times$ more tech-tree milestones and collecting $2.3\times$ more unique items than the Voyager baseline. Furthermore, in fully \textit{collaborative} settings, we find that the performance of two underachieving agents improves with more communication rounds, echoing the Condorcet Jury Theorem. MindForge agents demonstrate sophisticated behaviors, including expert-novice knowledge transfer, collaborative problem solving, and adaptation to out-of-distribution tasks through accumulated cultural experiences.
format Preprint
id arxiv_https___arxiv_org_abs_2411_12977
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle MindForge: Empowering Embodied Agents with Theory of Mind for Lifelong Cultural Learning
Lică, Mircea
Shirekar, Ojas
Colle, Baptiste
Raman, Chirag
Artificial Intelligence
Computation and Language
Embodied agents powered by large language models (LLMs), such as Voyager, promise open-ended competence in worlds such as Minecraft. However, when powered by open-weight LLMs they still falter on elementary tasks after domain-specific fine-tuning. We propose MindForge, a generative-agent framework for cultural lifelong learning through explicit perspective taking. We introduce three key innovations: (1) a structured theory of mind representation linking percepts, beliefs, desires, and actions; (2) natural inter-agent communication; and (3) a multi-component memory system. Following the cultural learning framework, we test MindForge in both instructive and collaborative settings within Minecraft. In an instructive setting with GPT-4, MindForge agents powered by open-weight LLMs significantly outperform their Voyager counterparts in basic tasks yielding $3\times$ more tech-tree milestones and collecting $2.3\times$ more unique items than the Voyager baseline. Furthermore, in fully \textit{collaborative} settings, we find that the performance of two underachieving agents improves with more communication rounds, echoing the Condorcet Jury Theorem. MindForge agents demonstrate sophisticated behaviors, including expert-novice knowledge transfer, collaborative problem solving, and adaptation to out-of-distribution tasks through accumulated cultural experiences.
title MindForge: Empowering Embodied Agents with Theory of Mind for Lifelong Cultural Learning
topic Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2411.12977