On the Generalization Gap in Reparameterizable Reinforcement Learning.
Huan WangStephan ZhengCaiming XiongRichard SocherPublished in: ICML (2019)
Keyphrases
- reinforcement learning
- function approximation
- learning algorithm
- state space
- databases
- model free
- supervised learning
- control problems
- dynamic programming
- multi agent
- reinforcement learning algorithms
- optimal control
- learning agents
- stochastic approximation
- partially observable
- real time
- transition model
- temporal difference
- transfer learning
- optimal policy
- artificial neural networks
- expert systems
- multiscale
- social networks