Over-communicate no more: Situated RL agents learn concise communication protocols.
Aleksandra KalinowskaElnaz DavoodiFlorian StrubKory W. MathewsonIvana KajicMichael BowlingTodd D. MurpheyPatrick M. PilarskiPublished in: CoRR (2022)
Keyphrases
- communication protocols
- communication protocol
- learning agent
- multi agent
- learning agents
- multi agent systems
- reinforcement learning agents
- reinforcement learning
- dynamic environments
- multiagent systems
- autonomous agents
- cooperative
- intelligent agents
- complex domains
- agent receives
- multiple agents
- communication networks
- action selection
- agent learns
- single agent
- temporal difference
- software agents
- mobile agents
- state space
- complex environments
- function approximation
- exchange information
- autonomous entities
- robocup soccer
- neural network