Least auxiliary loss-functions with impact growth adaptation (Laliga) for convolutional neural networks.
Mohammad Mahdi BejaniMehdi GhateePublished in: Neurocomputing (2021)
Keyphrases
- loss function
- convolutional neural networks
- pairwise
- logistic regression
- learning to rank
- loss minimization
- squared error
- risk minimization
- convolutional network
- empirical risk
- learning models
- support vector
- hinge loss
- pairwise constraints
- reproducing kernel hilbert space
- update rules
- learning to rank algorithms
- convex loss functions
- objective function