Successor Uncertainties: Exploration and Uncertainty in Temporal Difference Learning.
David JanzJiri HronPrzemyslaw MazurKatja HofmannJosé Miguel Hernández-LobatoSebastian TschiatschekPublished in: NeurIPS (2019)
Keyphrases
- temporal difference learning
- fixed point
- function approximation
- reinforcement learning
- evaluation function
- game playing
- approximate value iteration
- temporal difference
- reinforcement learning algorithms
- policy iteration
- neural network
- markov decision process
- function approximators
- action selection
- gaussian process
- monte carlo
- sufficient conditions
- state space
- training data