True Gradient-Based Training of Deep Binary Activated Neural Networks Via Continuous Binarization.
Charbel SakrJungwook ChoiZhuo WangKailash GopalakrishnanNaresh R. ShanbhagPublished in: ICASSP (2018)
Keyphrases
- neural network
- training process
- training algorithm
- feedforward neural networks
- pattern recognition
- neural network training
- feed forward neural networks
- error back propagation
- training patterns
- multi layer perceptron
- back propagation
- document images
- training set
- test set
- supervised learning
- artificial neural networks
- continuous valued
- deep architectures
- machine learning
- fuzzy systems
- data sets
- backpropagation algorithm
- genetic algorithm
- image processing
- training data
- activation function
- preprocessing
- piecewise linear
- fuzzy neural network
- multi layer
- recurrent neural networks
- fault diagnosis