Results of the NeurIPS 2023 Neural MMO Competition on Multi-task Reinforcement Learning

TL;DR

The paper reports on the NeurIPS 2023 Neural MMO Competition, highlighting top-performing goal-conditional policies that generalize across unseen tasks, maps, and opponents, with open-sourced code and results.

Contribution

It introduces a competitive benchmark for multi-task reinforcement learning in a complex MMO environment and provides the top solutions and their performance improvements.

Findings

Top solution scored 4x higher than baseline

Achieved generalization to unseen tasks, maps, and opponents

Open-sourced all code and policies

Abstract

We present the results of the NeurIPS 2023 Neural MMO Competition, which attracted over 200 participants and submissions. Participants trained goal-conditional policies that generalize to tasks, maps, and opponents never seen during training. The top solution achieved a score 4x higher than our baseline within 8 hours of training on a single 4090 GPU. We open-source everything relating to Neural MMO and the competition under the MIT license, including the policy weights and training code for our baseline and for the top submissions.