IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures.
Lasse EspeholtHubert SoyerRémi MunosKaren SimonyanVolodymyr MnihTom WardYotam DoronVlad FiroiuTim HarleyIain DunningShane LeggKoray KavukcuogluPublished in: ICML (2018)
Keyphrases
- scalable distributed
- reinforcement learning
- learning process
- learning environment
- language learning
- relative importance
- learning activities
- file system
- learning materials
- partially observable domains
- deep learning
- function approximation
- neural network
- markov decision processes
- optimal policy
- state space
- multi agent
- e learning
- learning objects
- weighted sum
- learning classifier systems
- subject matter
- order statistics
- bayesian networks
- active exploration