Training a feed-forward network by feeding gradients forward rather than by back-propagation of errors.
Roelof K. BrouwerPublished in: Neurocomputing (1997)
Keyphrases
- feed forward
- back propagation
- multilayer neural network
- recurrent networks
- output layer
- training algorithm
- feed forward neural networks
- neural network
- connection weights
- artificial neural networks
- feedforward neural networks
- backpropagation algorithm
- bp network
- neural nets
- hidden layer
- radial basis
- recurrent neural networks
- multilayer perceptron neural network
- error back propagation
- spiking neural networks
- bp neural network
- bp algorithm
- multilayer perceptron
- activation function
- fuzzy logic
- visual cortex
- learning algorithm
- levenberg marquardt
- hidden neurons
- biologically plausible
- wavelet neural network
- rbf network
- cascade correlation
- gradient vector
- spiking neurons
- learning rules
- training speed
- multi layer neural network
- multi layer
- network structure
- training set
- image sequences
- knowledge base
- decision making
- artificial intelligence
- genetic algorithm