Strategies for Affect-Controlled Action-Selection in Soar-RL.
Eric HogewoningJoost BroekensJeroen EggermontErnst G. P. BovenkampPublished in: IWINAC (2) (2007)
Keyphrases
- action selection
- reinforcement learning
- basal ganglia
- robot soccer
- temporal difference
- action space
- human robot
- decision making
- continuous state and action spaces
- knowledge level
- action selection mechanism
- general intelligence
- learned knowledge
- reinforcement learning algorithms
- real world
- model free
- human brain
- state space
- computer vision
- machine learning