SAN-RL: combining spreading activation networks and reinforcement learning to learn configurable behaviors.
Daniel M. GainesDon Mitchell WilkesKanok KusumalnukoolSiripun ThongchaiKazuhiko KawamuraJohn H. WhitePublished in: Mobile Robots (2001)
Keyphrases
- reinforcement learning
- spreading activation
- semantic network
- function approximation
- function approximators
- learning agent
- reinforcement learning algorithms
- state space
- document retrieval
- semantic knowledge
- optimal policy
- markov decision processes
- model free
- learning algorithm
- temporal difference
- natural language text
- agent learns
- lateral inhibition
- rl algorithms
- domain knowledge
- learning process
- word sense disambiguation
- document collections
- wordnet
- information retrieval