Optimal filtering algorithms for fast learning in feedforward neural networks

Neural Networks

Volumn 5, Issue 5, 1992, Pages 779-787

  Shah, Samir ^a   Palmieri, Francesco ^a   Datum, Michael ^a  

^a UNIVERSITY OF CONNECTICUT   (United States)

Author keywords

Fast back propagation; Feedforward networks; Kalman filtering; Neural networks; Nonlinear optimization; Optimal filtering; Pattern recognition; Supervised learning

Indexed keywords

ALGORITHMS; IDENTIFICATION (CONTROL SYSTEMS); KALMAN FILTERING; LEARNING SYSTEMS; LINEARIZATION; OPTIMIZATION; PATTERN RECOGNITION; PERFORMANCE; SIGNAL FILTERING AND PREDICTION;

FAST BACK PROPAGATION; FEEDFORWARD NETWORKS; NONLINEAR OPTIMIZATION; OPTIMAL FILTERING; SUPERVISED LEARNING;

NEURAL NETWORKS;

EID: 0026923239     PISSN: 08936080     EISSN: None     Source Type: Journal    
DOI: 10.1016/S0893-6080(05)80139-X     Document Type: Article

References (16)

1. Fast learning of multilayer perceptron using recursive least squares technique
2. Increased rates of convergence through learning rate adaptation
3. Fast neural nets with Gram-Schmidt orthogonalization
4. Sound localization with a neural network trained with the Multiple Extended Kalman Algorithm
5. Fast training of multilayer perceptrons using multilinear parameterization
6. MEKA—A fast, local algorithm for training feedforward neural networks
7. Decoupled Extended Kalman Filter training of feedforward layered networks
8. Rescaling of variables in Back Propagation learning
9. Learning internal representation by error propagation
10. Backpropagation improvements based on heuristic arguments
11. A fast algorithm for neural networks
12. Training feed-forward networks with the extended Kalman filter
13. Learning algorithms for connectionist networks: Applied gradient methods of non-linear optimization