Abstract | ||
---|---|---|
We study the emergence of cooperative behaviors in reinforcement learning agents by introducing a challenging competitive multi-agent soccer environment with continuous simulated physics. We demonstrate that decentralized, population-based training with co-play can lead to a progression in agentsu0027 behaviors: from random, to simple ball chasing, and finally showing evidence of cooperation. Our study highlights several of the challenges encountered in large scale multi-agent training in continuous control. In particular, we demonstrate that the automatic optimization of simple shaping rewards, not themselves conducive to co-operative behavior, can lead to long-horizon team behavior. We further apply an evaluation scheme, grounded by game theoretic principals, that can assess agent performance in the absence of pre-defined evaluation tasks or human baselines. |
Year | Venue | Field |
---|---|---|
2019 | ICLR | Computer science,Human–computer interaction,Artificial intelligence,Machine learning |
DocType | Volume | Citations |
Journal | abs/1902.07151 | 4 |
PageRank | References | Authors |
0.37 | 34 | 6 |
Name | Order | Citations | PageRank |
---|---|---|---|
Siqi Liu | 1 | 55 | 4.94 |
Guy Lever | 2 | 108 | 7.07 |
Josh S. Merel | 3 | 143 | 11.34 |
Saran Tunyasuvunakool | 4 | 10 | 2.14 |
Nicolas Heess | 5 | 1762 | 94.77 |
Thore Graepel | 6 | 4211 | 242.71 |