MACTAS: Self-Attention-Based Inter-Agent Communication in Multi-Agent Reinforcement Learning with Action-Value Function Decomposition
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866908782956642304 |
|---|---|
| author | Wojtala, Maciej Stefańczyk, Bogusz Bogucki, Dominik Lepak, Łukasz Strykowski, Jakub Wawrzyński, Paweł |
| author_facet | Wojtala, Maciej Stefańczyk, Bogusz Bogucki, Dominik Lepak, Łukasz Strykowski, Jakub Wawrzyński, Paweł |
| contents | Communication is essential for the collective execution of complex tasks by human agents, motivating interest in communication mechanisms for multi-agent reinforcement learning (MARL). However, existing communication protocols in MARL are often complex and non-differentiable. In this work, we introduce a self-attention-based communication method that exchanges information between the agents in MARL. Our proposed approach is fully differentiable, allowing agents to learn to generate messages in a reward-driven manner. The method can be seamlessly integrated with any action-value function decomposition algorithm and can be viewed as an orthogonal extension of such decompositions. Notably, it includes a fixed number of trainable parameters, independent of the number of agents, which makes it scalable to large systems. Experimental results on the SMACv2 benchmark demonstrate the effectiveness of our approach, which achieves state-of-the-art performance on a number of maps. makes it scalable to large systems. Experimental results on the SMACv2 benchmark demonstrate the effectiveness of our approach, which achieves state-of-the-art performance on a number of maps. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_13661 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MACTAS: Self-Attention-Based Inter-Agent Communication in Multi-Agent Reinforcement Learning with Action-Value Function Decomposition Wojtala, Maciej Stefańczyk, Bogusz Bogucki, Dominik Lepak, Łukasz Strykowski, Jakub Wawrzyński, Paweł Machine Learning Multiagent Systems Communication is essential for the collective execution of complex tasks by human agents, motivating interest in communication mechanisms for multi-agent reinforcement learning (MARL). However, existing communication protocols in MARL are often complex and non-differentiable. In this work, we introduce a self-attention-based communication method that exchanges information between the agents in MARL. Our proposed approach is fully differentiable, allowing agents to learn to generate messages in a reward-driven manner. The method can be seamlessly integrated with any action-value function decomposition algorithm and can be viewed as an orthogonal extension of such decompositions. Notably, it includes a fixed number of trainable parameters, independent of the number of agents, which makes it scalable to large systems. Experimental results on the SMACv2 benchmark demonstrate the effectiveness of our approach, which achieves state-of-the-art performance on a number of maps. makes it scalable to large systems. Experimental results on the SMACv2 benchmark demonstrate the effectiveness of our approach, which achieves state-of-the-art performance on a number of maps. |
| title | MACTAS: Self-Attention-Based Inter-Agent Communication in Multi-Agent Reinforcement Learning with Action-Value Function Decomposition |
| topic | Machine Learning Multiagent Systems |
| url | https://arxiv.org/abs/2508.13661 |