The Neural MMO Platform for Massively Multiagent Research
Neural MMO is a computationally accessible research platform that combines large agent populations, long time horizons, open-ended tasks, and modular game systems. Existing environments feature subsets of these properties, but Neural MMO is the first to combine them all. We present Neural MMO as fre...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Published: |
Massachusetts Institute of Technology
2023
|
| Online Access: | https://hdl.handle.net/1721.1/151250 |
| _version_ | 1826191714501525504 |
|---|---|
| author | Suarez, Joseph |
| author2 | Isola, Philip |
| author_facet | Isola, Philip Suarez, Joseph |
| author_sort | Suarez, Joseph |
| collection | MIT |
| description | Neural MMO is a computationally accessible research platform that combines large agent populations, long time horizons, open-ended tasks, and modular game systems. Existing environments feature subsets of these properties, but Neural MMO is the first to combine them all. We present Neural MMO as free and open source software with active support, ongoing development, documentation, and additional training, logging, and visualization tools to help users adapt to this new setting. Initial baselines on the platform demonstrate that agents trained in large populations explore more and learn a progression of skills. We raise other more difficult problems such as many-team cooperation as open research questions which Neural MMO is well-suited to answer. Finally, we discuss current limitations of the platform, potential mitigations, and plans for continued development. |
| first_indexed | 2024-09-23T09:00:13Z |
| format | Thesis |
| id | mit-1721.1/151250 |
| institution | Massachusetts Institute of Technology |
| last_indexed | 2024-09-23T09:00:13Z |
| publishDate | 2023 |
| publisher | Massachusetts Institute of Technology |
| record_format | dspace |
| spelling | mit-1721.1/1512502023-08-01T04:14:59Z The Neural MMO Platform for Massively Multiagent Research Suarez, Joseph Isola, Philip Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Neural MMO is a computationally accessible research platform that combines large agent populations, long time horizons, open-ended tasks, and modular game systems. Existing environments feature subsets of these properties, but Neural MMO is the first to combine them all. We present Neural MMO as free and open source software with active support, ongoing development, documentation, and additional training, logging, and visualization tools to help users adapt to this new setting. Initial baselines on the platform demonstrate that agents trained in large populations explore more and learn a progression of skills. We raise other more difficult problems such as many-team cooperation as open research questions which Neural MMO is well-suited to answer. Finally, we discuss current limitations of the platform, potential mitigations, and plans for continued development. S.M. 2023-07-31T19:26:01Z 2023-07-31T19:26:01Z 2023-06 2023-07-13T14:29:57.003Z Thesis https://hdl.handle.net/1721.1/151250 Attribution 4.0 International (CC BY 4.0) Copyright retained by author(s) https://creativecommons.org/licenses/by/4.0/ application/pdf Massachusetts Institute of Technology |
| spellingShingle | Suarez, Joseph The Neural MMO Platform for Massively Multiagent Research |
| title | The Neural MMO Platform for Massively Multiagent Research |
| title_full | The Neural MMO Platform for Massively Multiagent Research |
| title_fullStr | The Neural MMO Platform for Massively Multiagent Research |
| title_full_unstemmed | The Neural MMO Platform for Massively Multiagent Research |
| title_short | The Neural MMO Platform for Massively Multiagent Research |
| title_sort | neural mmo platform for massively multiagent research |
| url | https://hdl.handle.net/1721.1/151250 |
| work_keys_str_mv | AT suarezjoseph theneuralmmoplatformformassivelymultiagentresearch AT suarezjoseph neuralmmoplatformformassivelymultiagentresearch |