ClusterComm: Discrete Communication in Decentralized MARL Using Internal Representation Clustering
Robert Müller, Hasan Turalic, Thomy Phan, Michael Kölle, Jonas Nüßlein, Claudia Linnhoff-Popien
2024
Abstract
In the realm of Multi-Agent Reinforcement Learning (MARL), prevailing approaches exhibit shortcomings in aligning with human learning, robustness, and scalability. Addressing this, we introduce ClusterComm, a fully decentralized MARL framework where agents communicate discretely without a central control unit. ClusterComm utilizes Mini-Batch-K-Means clustering on the last hidden layer’s activations of an agent’s policy network, translating them into discrete messages. This approach outperforms no communication and competes favorably with unbounded, continuous communication and hence poses a simple yet effective strategy for enhancing collaborative task-solving in MARL.
DownloadPaper Citation
in Harvard Style
Müller R., Turalic H., Phan T., Kölle M., Nüßlein J. and Linnhoff-Popien C. (2024). ClusterComm: Discrete Communication in Decentralized MARL Using Internal Representation Clustering. In Proceedings of the 16th International Conference on Agents and Artificial Intelligence - Volume 1: ICAART; ISBN 978-989-758-680-4, SciTePress, pages 305-312. DOI: 10.5220/0012384300003636
in Bibtex Style
@conference{icaart24,
author={Robert Müller and Hasan Turalic and Thomy Phan and Michael Kölle and Jonas Nüßlein and Claudia Linnhoff-Popien},
title={ClusterComm: Discrete Communication in Decentralized MARL Using Internal Representation Clustering},
booktitle={Proceedings of the 16th International Conference on Agents and Artificial Intelligence - Volume 1: ICAART},
year={2024},
pages={305-312},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012384300003636},
isbn={978-989-758-680-4},
}
in EndNote Style
TY - CONF
JO - Proceedings of the 16th International Conference on Agents and Artificial Intelligence - Volume 1: ICAART
TI - ClusterComm: Discrete Communication in Decentralized MARL Using Internal Representation Clustering
SN - 978-989-758-680-4
AU - Müller R.
AU - Turalic H.
AU - Phan T.
AU - Kölle M.
AU - Nüßlein J.
AU - Linnhoff-Popien C.
PY - 2024
SP - 305
EP - 312
DO - 10.5220/0012384300003636
PB - SciTePress