Linearly Constrained Weights: Reducing Activation Shift for Faster Training of Neural Networks.
Takuro KutsunaPublished in: ECML/PKDD (2) (2019)
Keyphrases
- neural network
- linearly constrained
- training process
- training algorithm
- feed forward neural networks
- feedforward neural networks
- multi layer perceptron
- linear constraints
- back propagation
- linearly combined
- fuzzy logic
- feed forward
- activation function
- artificial neural networks
- variational inequalities
- backpropagation algorithm
- upper bound
- pairwise
- training set
- neural network structure
- training data
- image processing