ClusterComm: Discrete Communication in Decentralized MARL using Internal Representation Clustering
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914633724461056 |
|---|---|
| author | Müller, Robert Turalic, Hasan Phan, Thomy Kölle, Michael Nüßlein, Jonas Linnhoff-Popien, Claudia |
| author_facet | Müller, Robert Turalic, Hasan Phan, Thomy Kölle, Michael Nüßlein, Jonas Linnhoff-Popien, Claudia |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2401_03504 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | ClusterComm: Discrete Communication in Decentralized MARL using Internal Representation Clustering Müller, Robert Turalic, Hasan Phan, Thomy Kölle, Michael Nüßlein, Jonas Linnhoff-Popien, Claudia Artificial Intelligence 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. |
| title | ClusterComm: Discrete Communication in Decentralized MARL using Internal Representation Clustering |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2401.03504 |