DigiRL: Training In-The-Wild Device-Control Agents with Autonomous Reinforcement Learning.
Hao BaiYifei ZhouMert CemriJiayi PanAlane SuhrSergey LevineAviral KumarPublished in: CoRR (2024)
Keyphrases
- reinforcement learning
- decentralized control
- multi agent
- adjustable autonomy
- multi agent systems
- autonomous systems
- action selection
- cooperative
- optimal control
- multiagent systems
- learning agents
- control problems
- learning capabilities
- autonomous agents
- robot control
- distributed control
- intelligent agents
- cooperative behavior
- control policy
- autonomous learning
- multiple agents
- agents and multi agent systems
- evolutionary learning
- control system
- robotic systems
- data acquisition
- multiagent learning
- multi agent environments
- learning algorithm
- control strategies
- training process
- single agent
- autonomous vehicles
- model free
- state space
- multi agent learning
- mobile agents
- decision theoretic
- multi agent reinforcement learning
- reinforcement learning algorithms
- neural network
- reinforcement learning agents
- decision making
- function approximation
- robocup soccer
- agent receives
- markov decision processes
- software agents
- control strategy
- partial observability
- agent behavior
- coalition formation
- temporal difference