The reviewed record of science sign in
Pith

arxiv: 2501.01266 · v1 · pith:Q2P7S2Z4 · submitted 2025-01-02 · cs.MA · cs.AI

PIMAEX: Multi-Agent Exploration through Peer Incentivization

Reviewed by Pith T0 review T1 audit T2 compute T3 formal T4 kernel pith:Q2P7S2Z4record.jsonopen to challenge →

classification cs.MA cs.AI
keywords textitexplorationmulti-agentpimaexrewardagentsenvironmentinfluence
0
0 comments X
read the original abstract

While exploration in single-agent reinforcement learning has been studied extensively in recent years, considerably less work has focused on its counterpart in multi-agent reinforcement learning. To address this issue, this work proposes a peer-incentivized reward function inspired by previous research on intrinsic curiosity and influence-based rewards. The \textit{PIMAEX} reward, short for Peer-Incentivized Multi-Agent Exploration, aims to improve exploration in the multi-agent setting by encouraging agents to exert influence over each other to increase the likelihood of encountering novel states. We evaluate the \textit{PIMAEX} reward in conjunction with \textit{PIMAEX-Communication}, a multi-agent training algorithm that employs a communication channel for agents to influence one another. The evaluation is conducted in the \textit{Consume/Explore} environment, a partially observable environment with deceptive rewards, specifically designed to challenge the exploration vs.\ exploitation dilemma and the credit-assignment problem. The results empirically demonstrate that agents using the \textit{PIMAEX} reward with \textit{PIMAEX-Communication} outperform those that do not.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.