Partitioning sparse deep neural networks for scalable training and inference.
Gunduz Vehbi DemirciHakan FerhatosmanogluPublished in: CoRR (2021)
Keyphrases
- neural network
- training process
- training algorithm
- structured prediction
- feedforward neural networks
- backpropagation algorithm
- deep architectures
- back propagation
- feed forward neural networks
- error back propagation
- inference process
- multi layer perceptron
- pattern recognition
- artificial neural networks
- data sets
- genetic algorithm
- neural network structure
- bayesian networks
- training set
- training samples
- avoid overfitting
- neural network training
- feed forward
- reduced set
- multilayer perceptron
- radial basis function network
- recurrent neural networks
- high dimensional
- training examples
- sparse data
- fuzzy logic
- partitioning algorithm
- graph partitioning
- online learning
- neural network model
- fault diagnosis
- multi layer
- test set