A partially recurrent gating network approach to learning action selection by reinforcement.
R. Mark RylattChris A. CzarneckiTom W. RoutenPublished in: ICNN (1997)
Keyphrases
- action selection
- reinforcement learning
- learning algorithm
- recurrent networks
- connectionist networks
- learning process
- continuous state and action spaces
- computer networks
- biologically inspired
- network model
- online learning
- basal ganglia
- data mining
- learning systems
- network structure
- learning tasks
- learning rules
- robot soccer
- supervised learning
- prior knowledge