Improving the State Space Organization of Untrained Recurrent Networks.
Michal CernanskýMatej MakulaLubica BenuskováPublished in: ICONIP (1) (2008)
Keyphrases
- recurrent networks
- state space
- recurrent neural networks
- dynamical systems
- reinforcement learning
- feed forward
- neural network
- state variables
- biologically inspired
- heuristic search
- optimal policy
- planning problems
- markov decision processes
- particle filter
- markov chain
- search space
- real time
- knowledge management
- dynamic programming
- spatio temporal
- optimal solution
- systems development
- real world
- continuous state spaces