Stochastic smoothing of the top-K calibrated hinge loss for deep imbalanced classification.
Camille GarcinMaximilien ServajeanAlexis JolyJoseph SalmonPublished in: CoRR (2022)
Keyphrases
- hinge loss
- loss function
- decision boundary
- classification accuracy
- decision trees
- convex optimization
- support vector machine
- pattern classification
- potential functions
- machine learning
- risk minimization
- support vector
- feature extraction
- multi class
- text classification
- svm classifier
- image classification
- feature space
- reproducing kernel hilbert space