Analysing factorizations of action-value networks for cooperative multi-agent reinforcement learning

Analysing factorizations of action-value networks for cooperative multi-agent reinforcement learning

Castellini, Jacopo (University of Liverpool)
Oliehoek, F.A. (TU Delft Interactive Intelligence)
Savani, Rahul (University of Liverpool)
Whiteson, Shimon (University of Oxford)

2021

Recent years have seen the application of deep reinforcement learning techniques to cooperative multi-agent systems, with great empirical success. However, given the lack of theoretical insight, it remains unclear what the employed neural networks are learning, or how we should enhance their learning power to address the problems on which they fail. In this work, we empirically investigate the learning power of various network architectures on a series of one-shot games. Despite their simplicity, these games capture many of the crucial problems that arise in the multi-agent setting, such as an exponential number of joint actions or the lack of an explicit coordination mechanism. Our results extend those in Castellini et al. (Proceedings of the 18th International Conference on Autonomous Agents and MultiAgent Systems, AAMAS’19.International Foundation for Autonomous Agents and Multiagent Systems, pp 1862–1864, 2019) and quantify how well various approaches can represent the requisite value functions, and help us identify the reasons that can impede good performance, like sparsity of the values or too tight coordination requirements.

Action-value representation
Decision-making
Multi-agent systems
Neural networks
One-shot games

http://resolver.tudelft.nl/uuid:49288083-b5f9-431f-8371-47dc4b789134

https://doi.org/10.1007/s10458-021-09506-w

1387-2532

Autonomous Agents and Multi-Agent Systems, 35 (2), 1-53

Institutional Repository

journal article

© 2021 Jacopo Castellini, F.A. Oliehoek, Rahul Savani, Shimon Whiteson