Improving generalization to new environments and removing catastrophic forgetting in Reinforcement Learning by using an eco-system of agents.
Olivier MoulinFrancois Vincent-LavetPaul W. G. ElbersMark HoogendoornPublished in: WI/IAT (2022)
Keyphrases
- reinforcement learning
- multi agent environments
- multi agent
- multi agent systems
- learning agents
- autonomous agents
- dynamic environments
- action selection
- multiagent systems
- highly dynamic
- complex environments
- single agent
- multi agent reinforcement learning
- multiple agents
- multiagent learning
- reinforcement learning algorithms
- intelligent agents
- open systems
- reinforcement learning agents
- agent receives
- function approximation
- software agents
- agent model
- open environments
- learning agent
- cooperative
- artificial agents
- decision making
- state space
- optimal policy
- agent systems
- robocup soccer
- cognitive agents
- evolutionary learning
- resource allocation
- markov decision process
- autonomous entities
- incremental learning
- agent behavior
- mobile agents
- real world
- temporal difference
- state abstraction
- markov decision processes
- partial observability
- optimal control
- learned knowledge
- agent technology
- function approximators
- model free