MES-Loss: Mutually equidistant separation metric learning loss function.
Yasser BoutalebCatherine SoladiéNam-Duong DuongAmine KaceteJérôme RoyanRenaud SéguierPublished in: Pattern Recognit. Lett. (2023)
Keyphrases
- metric learning
- loss function
- pairwise
- hinge loss
- convex loss functions
- semi supervised
- pairwise constraints
- distance metric learning
- distance metric
- risk minimization
- support vector
- kernel matrix
- distance function
- learning tasks
- multi task
- feature space
- dimensionality reduction
- mahalanobis metric
- reproducing kernel hilbert space
- semi supervised learning
- markov random field
- similarity measure
- multi class
- binary classification
- unsupervised learning
- feature extraction
- line search
- image segmentation
- learning algorithm
- data mining