“Neural MMO 2.0: A Massively Multi-Task Addition to Massively Multi-Agent Learning”, Joseph Suárez, Phillip Isola, Kyoung Whan Choe, David Bloomin, Hao Xiang Li, Nikhil Pinnaparaju, Nishaanth Kanna, Daniel Scott, Ryan Sullivan, Rose S. Shuman, Lucas de Alcântara, Herbie Bradley, Louis Castricato, Kirsty You, Yuhao Jiang, Qimai Li, Jiaxin Chen, Xiaolong Zhu2023-11-07 (MARL)⁠:

Neural MMO 2.0 [1.0] is a massively multi-agent environment for reinforcement learning research. The key feature of this new version is a flexible task system that allows users to define a broad range of objectives and reward signals.

We challenge researchers to train agents capable of generalizing to tasks, maps, and opponents never seen during training. Neural MMO features procedurally generated maps with 128 agents in the standard setting and support for more. Version 2.0 is a complete rewrite of its predecessor with 3x improved performance and compatibility with Clean RL.

We release the platform as free and open-source software with comprehensive documentation available at our homepage and an active community Discord.

To spark initial research on this new platform, we are concurrently running a competition at NeurIPS 2023.