The NeurIPS 2022 Neural MMO Challenge: A Massively Multiagent Competition with Specialization and Trade

Enhong Liu, Joseph Suarez, Chenhui You, Bo Wu, Bingcheng Chen, Jun Hu, Jiaxin Chen, Xiaolong Zhu, Clare Zhu, Julian Togelius, Sharada Mohanty, Weijun Hong, Rui Du, Yibing Zhang, Qinwen Wang, Xinhang Li, Zheng Yuan, Xiang Li, Yuejia Huang, Kun Zhang, Hanhui Yang, Shiqi Tang, Phillip Isola
Proceedings of the NeurIPS 2022 Competitions Track, PMLR 220:18-34, 2022.

Abstract

In this paper, we present the results of the NeurIPS-2022 Neural MMO Challenge, which attracted 500 participants and received over 1,600 submissions. Like the previous IJCAI-2022 Neural MMO Challenge, it involved agents from 16 populations surviving in procedurally generated worlds by collecting resources and defeating opponents. This year’s competition runs on the latest v1.6 Neural MMO, which introduces new equipment, combat, trading, and a better scoring system. These elements combine to pose additional robustness and generalization challenges not present in previous competitions. This paper summarizes the design and results of the challenge, explores the potential of this environment as a benchmark for learning methods, and presents some practical reinforcement learning training approaches for complex tasks with sparse rewards. Additionally, we have open-sourced our baselines, including environment wrappers, benchmarks, and visualization tools for future research.

Cite this Paper

BibTeX
@InProceedings{pmlr-v220-liu23a, title = {The NeurIPS 2022 Neural MMO Challenge: A Massively Multiagent Competition with Specialization and Trade}, author = {Liu, Enhong and Suarez, Joseph and You, Chenhui and Wu, Bo and Chen, Bingcheng and Hu, Jun and Chen, Jiaxin and Zhu, Xiaolong and Zhu, Clare and Togelius, Julian and Mohanty, Sharada and Hong, Weijun and Du, Rui and Zhang, Yibing and Wang, Qinwen and Li, Xinhang and Yuan, Zheng and Li, Xiang and Huang, Yuejia and Zhang, Kun and Yang, Hanhui and Tang, Shiqi and Isola, Phillip}, booktitle = {Proceedings of the NeurIPS 2022 Competitions Track}, pages = {18--34}, year = {2022}, editor = {Ciccone, Marco and Stolovitzky, Gustavo and Albrecht, Jacob}, volume = {220}, series = {Proceedings of Machine Learning Research}, month = {28 Nov--09 Dec}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v220/liu23a/liu23a.pdf}, url = {https://proceedings.mlr.press/v220/liu23a.html}, abstract = {In this paper, we present the results of the NeurIPS-2022 Neural MMO Challenge, which attracted 500 participants and received over 1,600 submissions. Like the previous IJCAI-2022 Neural MMO Challenge, it involved agents from 16 populations surviving in procedurally generated worlds by collecting resources and defeating opponents. This year’s competition runs on the latest v1.6 Neural MMO, which introduces new equipment, combat, trading, and a better scoring system. These elements combine to pose additional robustness and generalization challenges not present in previous competitions. This paper summarizes the design and results of the challenge, explores the potential of this environment as a benchmark for learning methods, and presents some practical reinforcement learning training approaches for complex tasks with sparse rewards. Additionally, we have open-sourced our baselines, including environment wrappers, benchmarks, and visualization tools for future research.} }
Endnote
%0 Conference Paper %T The NeurIPS 2022 Neural MMO Challenge: A Massively Multiagent Competition with Specialization and Trade %A Enhong Liu %A Joseph Suarez %A Chenhui You %A Bo Wu %A Bingcheng Chen %A Jun Hu %A Jiaxin Chen %A Xiaolong Zhu %A Clare Zhu %A Julian Togelius %A Sharada Mohanty %A Weijun Hong %A Rui Du %A Yibing Zhang %A Qinwen Wang %A Xinhang Li %A Zheng Yuan %A Xiang Li %A Yuejia Huang %A Kun Zhang %A Hanhui Yang %A Shiqi Tang %A Phillip Isola %B Proceedings of the NeurIPS 2022 Competitions Track %C Proceedings of Machine Learning Research %D 2022 %E Marco Ciccone %E Gustavo Stolovitzky %E Jacob Albrecht %F pmlr-v220-liu23a %I PMLR %P 18--34 %U https://proceedings.mlr.press/v220/liu23a.html %V 220 %X In this paper, we present the results of the NeurIPS-2022 Neural MMO Challenge, which attracted 500 participants and received over 1,600 submissions. Like the previous IJCAI-2022 Neural MMO Challenge, it involved agents from 16 populations surviving in procedurally generated worlds by collecting resources and defeating opponents. This year’s competition runs on the latest v1.6 Neural MMO, which introduces new equipment, combat, trading, and a better scoring system. These elements combine to pose additional robustness and generalization challenges not present in previous competitions. This paper summarizes the design and results of the challenge, explores the potential of this environment as a benchmark for learning methods, and presents some practical reinforcement learning training approaches for complex tasks with sparse rewards. Additionally, we have open-sourced our baselines, including environment wrappers, benchmarks, and visualization tools for future research.
APA
Liu, E., Suarez, J., You, C., Wu, B., Chen, B., Hu, J., Chen, J., Zhu, X., Zhu, C., Togelius, J., Mohanty, S., Hong, W., Du, R., Zhang, Y., Wang, Q., Li, X., Yuan, Z., Li, X., Huang, Y., Zhang, K., Yang, H., Tang, S. & Isola, P.. (2022). The NeurIPS 2022 Neural MMO Challenge: A Massively Multiagent Competition with Specialization and Trade. Proceedings of the NeurIPS 2022 Competitions Track, in Proceedings of Machine Learning Research 220:18-34 Available from https://proceedings.mlr.press/v220/liu23a.html.

Related Material