Applying Policy Iteration for Training Recurrent Neural Networks
Istvan SzitaAndrás LörinczPublished in: CoRR (2004)
Keyphrases
- recurrent neural networks
- policy iteration
- recurrent networks
- echo state networks
- feedforward neural networks
- markov decision processes
- feed forward
- model free
- reinforcement learning
- sample path
- neural network
- artificial neural networks
- least squares
- optimal policy
- fixed point
- reservoir computing
- training set
- finite state
- nonlinear dynamic systems
- temporal difference
- supervised learning
- average reward
- optimal control
- policy evaluation
- linear programming