Automata-Conditioned Cooperative Multi-Agent Reinforcement Learning
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866912686873247744 |
|---|---|
| author | Yalcinkaya, Beyazit Vazquez-Chanlatte, Marcell Shah, Ameesh Krasowski, Hanna Seshia, Sanjit A. |
| author_facet | Yalcinkaya, Beyazit Vazquez-Chanlatte, Marcell Shah, Ameesh Krasowski, Hanna Seshia, Sanjit A. |
| contents | We study the problem of learning multi-task, multi-agent policies for cooperative, temporal objectives, under centralized training, decentralized execution. In this setting, using automata to represent tasks enables the decomposition of complex tasks into simpler sub-tasks that can be assigned to agents. However, existing approaches remain sample-inefficient and are limited to the single-task case. In this work, we present Automata-Conditioned Cooperative Multi-Agent Reinforcement Learning (ACC-MARL), a framework for learning task-conditioned, decentralized team policies. We identify the main challenges to ACC-MARL's feasibility in practice, propose solutions, and prove the correctness of our approach. We further show that the value functions of learned policies can be used to assign tasks optimally at test time. Experiments show emergent task-aware, multi-step coordination among agents, e.g., pressing a button to unlock a door, holding the door, and short-circuiting tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_02304 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Automata-Conditioned Cooperative Multi-Agent Reinforcement Learning Yalcinkaya, Beyazit Vazquez-Chanlatte, Marcell Shah, Ameesh Krasowski, Hanna Seshia, Sanjit A. Multiagent Systems Artificial Intelligence Computation and Language Formal Languages and Automata Theory Machine Learning We study the problem of learning multi-task, multi-agent policies for cooperative, temporal objectives, under centralized training, decentralized execution. In this setting, using automata to represent tasks enables the decomposition of complex tasks into simpler sub-tasks that can be assigned to agents. However, existing approaches remain sample-inefficient and are limited to the single-task case. In this work, we present Automata-Conditioned Cooperative Multi-Agent Reinforcement Learning (ACC-MARL), a framework for learning task-conditioned, decentralized team policies. We identify the main challenges to ACC-MARL's feasibility in practice, propose solutions, and prove the correctness of our approach. We further show that the value functions of learned policies can be used to assign tasks optimally at test time. Experiments show emergent task-aware, multi-step coordination among agents, e.g., pressing a button to unlock a door, holding the door, and short-circuiting tasks. |
| title | Automata-Conditioned Cooperative Multi-Agent Reinforcement Learning |
| topic | Multiagent Systems Artificial Intelligence Computation and Language Formal Languages and Automata Theory Machine Learning |
| url | https://arxiv.org/abs/2511.02304 |