A Dissection of Overfitting and Generalization in Continuous Reinforcement Learning.
Amy ZhangNicolas BallasJoelle PineauPublished in: CoRR (2018)
Keyphrases
- reinforcement learning
- action space
- function approximation
- continuous state and action spaces
- state space
- continuous domains
- robotic control
- continuous state spaces
- cross validation
- markov decision processes
- decision trees
- training error
- fitted q iteration
- neural network
- continuous functions
- continuous data
- optimal policy
- dynamic programming
- website
- robot control
- reinforcement learning algorithms
- temporal difference
- learning process
- multi agent
- learning algorithm