Gossip-based Actor-Learner Architectures for Deep Reinforcement Learning.
Mahmoud AssranJoshua RomoffNicolas BallasJoelle PineauMike RabbatPublished in: NeurIPS (2019)
Keyphrases
- reinforcement learning
- learning process
- active exploration
- function approximation
- learning environment
- state space
- learning materials
- model free
- learning algorithm
- e learning
- markov decision processes
- adaptive learning
- virtual learning environments
- multi agent reinforcement learning
- reinforcement learning algorithms
- learner model
- robotic control
- deep learning
- iterative learning
- language learning
- optimal control
- temporal difference learning
- function approximators
- temporal difference
- data sets
- optimal policy
- dynamic programming
- neural network