Towards a Better Understanding of Learning with Multiagent Teams
Abstract
While it has long been recognized that a team of individual learning agents can be greater than the sum of its parts, recent work has shown that larger teams are not necessarily more effective than smaller ones. In this paper, we study why and under which conditions certain team structures promote effective learning for a population of individual learning agents. We show that, depending on the environment, some team structures help agents learn to specialize into specific roles, resulting in more favorable global results. However, large teams create credit assignment challenges that reduce coordination, leading to large teams performing poorly compared to smaller ones. We support our conclusions with both theoretical analysis and empirical results.
Cite
Text
Radke et al. "Towards a Better Understanding of Learning with Multiagent Teams." International Joint Conference on Artificial Intelligence, 2023. doi:10.24963/IJCAI.2023/31Markdown
[Radke et al. "Towards a Better Understanding of Learning with Multiagent Teams." International Joint Conference on Artificial Intelligence, 2023.](https://mlanthology.org/ijcai/2023/radke2023ijcai-better/) doi:10.24963/IJCAI.2023/31BibTeX
@inproceedings{radke2023ijcai-better,
title = {{Towards a Better Understanding of Learning with Multiagent Teams}},
author = {Radke, David and Larson, Kate and Brecht, Tim and Tilbury, Kyle},
booktitle = {International Joint Conference on Artificial Intelligence},
year = {2023},
pages = {271-279},
doi = {10.24963/IJCAI.2023/31},
url = {https://mlanthology.org/ijcai/2023/radke2023ijcai-better/}
}