Contextual Behaviors and Internal Representations Acquired by Reinforcement Learning with a Recurrent Neural Network in a Continuous State and Action Space Task.
Hiroki UtsunomiyaKatsunari ShibataPublished in: ICONIP (2) (2008)
Keyphrases
- action space
- continuous state
- reinforcement learning
- state space
- markov decision processes
- real valued
- policy search
- control policies
- continuous state spaces
- state action
- stochastic processes
- action selection
- function approximators
- heuristic search
- reinforcement learning algorithms
- single agent
- continuous action
- function approximation
- neural network
- markov chain
- dynamic programming
- learning algorithm
- finite state
- dynamical systems
- optimal policy
- average reward
- higher order
- multi agent
- machine learning