ACE (Actor-Critic-Explorer) paradigm for reinforcement learning in basal ganglia: Highlighting the role of subthalamic and pallidal nuclei.
Denny JosephGaripelli GangadharV. Srinivasa ChakravarthyPublished in: Neurocomputing (2010)
Keyphrases
- basal ganglia
- actor critic
- action selection
- reinforcement learning
- temporal difference
- reinforcement learning algorithms
- function approximation
- policy gradient
- human brain
- approximate dynamic programming
- optimal control
- connectionist models
- gradient method
- policy iteration
- multi agent
- policy gradient methods
- neuro fuzzy
- markov decision processes
- model free
- state space
- natural actor critic
- machine learning
- learning algorithm
- temporal difference learning
- optimal policy
- function approximators
- average reward
- neural network
- rl algorithms
- control problems
- partially observable
- dynamic programming
- artificial intelligence
- sufficient conditions
- action space