Kernelized value function approximation for reinforcement learning.
Gavin TaylorRonald ParrPublished in: ICML (2009)
Keyphrases
- reinforcement learning
- temporal difference
- state space
- temporal difference learning
- approximate dynamic programming
- function approximation
- state action
- function approximators
- reinforcement learning algorithms
- markov games
- model free
- markov decision processes
- action selection
- basis functions
- partially observable
- control problems
- dynamic programming
- multi agent
- supervised learning
- policy iteration
- markov decision process
- feature selection
- continuous state
- reinforcement learning methods
- learning problems
- kernel function
- step size
- linear combination
- real robot
- optimal policy
- learning algorithm
- policy search