Shaping the State Space Landscape in Recurrent Networks.
Patrice Y. SimardJean Pierre RayszBernard VictorriPublished in: NIPS (1990)
Keyphrases
- recurrent networks
- state space
- recurrent neural networks
- reward shaping
- feed forward
- heuristic search
- biologically inspired
- neural network
- markov chain
- reinforcement learning
- orders of magnitude
- markov decision processes
- optimal policy
- dynamic programming
- reinforcement learning algorithms
- state variables
- planning problems
- search space
- markov decision process
- particle filter
- action space
- data sets
- goal state
- data mining
- dynamical systems
- complex domains
- markov decision problems
- high dimensional
- continuous state spaces
- partially observable
- spatio temporal
- belief state
- climate change
- pattern databases
- reward function
- fitness landscape