The Multiscale Structure of Neural Network Loss Functions: The Effect on Optimization and Origin.
Chao MaLei WuLexing YingPublished in: CoRR (2022)
Keyphrases
- loss function
- multiscale
- neural network
- loss minimization
- risk minimization
- pairwise
- learning to rank
- squared error
- image segmentation
- hinge loss
- back propagation
- support vector
- logistic regression
- image processing
- convex loss functions
- reproducing kernel hilbert space
- markov random field
- multi class
- artificial neural networks
- coordinate descent method