Uncertainty-driven Imagination for Continuous Deep Reinforcement Learning.
Gabriel KalweitJoschka BoedeckerPublished in: CoRL (2017)
Keyphrases
- reinforcement learning
- action space
- function approximation
- markov decision processes
- state space
- sequential decision problems
- partial observability
- fitted q iteration
- temporal difference
- data sets
- uncertain data
- continuous state spaces
- continuous state and action spaces
- data driven
- learning process
- multi agent
- dynamic programming
- reinforcement learning algorithms
- partially observable
- continuous domains
- control problems
- uncertain information
- stochastic approximation
- belief functions
- possibility theory
- piecewise linear
- conditional probabilities
- incomplete information
- transfer learning
- monte carlo
- least squares
- bayesian networks