Training Integrable Parameterizations of Deep Neural Networks in the Infinite-Width Limit.
Karl HajjarLénaïc ChizatChristophe GiraudPublished in: CoRR (2021)
Keyphrases
- neural network
- training process
- training algorithm
- feedforward neural networks
- back propagation
- feed forward neural networks
- backpropagation algorithm
- deep architectures
- multi layer perceptron
- pattern recognition
- hidden layer
- error back propagation
- neural network structure
- neural network training
- recurrent networks
- deep learning
- training set
- neural nets
- feed forward
- test set
- fuzzy logic
- expert systems
- genetic algorithm
- recurrent neural networks
- training patterns
- training samples