Generalising Multi-Agent Cooperation through Task-Agnostic Communication

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Jayalath, Dulhan, Morad, Steven, Prorok, Amanda
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916155851014144
author Jayalath, Dulhan
Morad, Steven
Prorok, Amanda
author_facet Jayalath, Dulhan
Morad, Steven
Prorok, Amanda
contents Existing communication methods for multi-agent reinforcement learning (MARL) in cooperative multi-robot problems are almost exclusively task-specific, training new communication strategies for each unique task. We address this inefficiency by introducing a communication strategy applicable to any task within a given environment. We pre-train the communication strategy without task-specific reward guidance in a self-supervised manner using a set autoencoder. Our objective is to learn a fixed-size latent Markov state from a variable number of agent observations. Under mild assumptions, we prove that policies using our latent representations are guaranteed to converge, and upper bound the value error introduced by our Markov state approximation. Our method enables seamless adaptation to novel tasks without fine-tuning the communication strategy, gracefully supports scaling to more agents than present during training, and detects out-of-distribution events in an environment. Empirical results on diverse MARL scenarios validate the effectiveness of our approach, surpassing task-specific communication strategies in unseen tasks. Our implementation of this work is available at https://github.com/proroklab/task-agnostic-comms.
format Preprint
id arxiv_https___arxiv_org_abs_2403_06750
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Generalising Multi-Agent Cooperation through Task-Agnostic Communication
Jayalath, Dulhan
Morad, Steven
Prorok, Amanda
Multiagent Systems
Machine Learning
Robotics
Existing communication methods for multi-agent reinforcement learning (MARL) in cooperative multi-robot problems are almost exclusively task-specific, training new communication strategies for each unique task. We address this inefficiency by introducing a communication strategy applicable to any task within a given environment. We pre-train the communication strategy without task-specific reward guidance in a self-supervised manner using a set autoencoder. Our objective is to learn a fixed-size latent Markov state from a variable number of agent observations. Under mild assumptions, we prove that policies using our latent representations are guaranteed to converge, and upper bound the value error introduced by our Markov state approximation. Our method enables seamless adaptation to novel tasks without fine-tuning the communication strategy, gracefully supports scaling to more agents than present during training, and detects out-of-distribution events in an environment. Empirical results on diverse MARL scenarios validate the effectiveness of our approach, surpassing task-specific communication strategies in unseen tasks. Our implementation of this work is available at https://github.com/proroklab/task-agnostic-comms.
title Generalising Multi-Agent Cooperation through Task-Agnostic Communication
topic Multiagent Systems
Machine Learning
Robotics
url https://arxiv.org/abs/2403.06750