On the interplay between data structure and loss function in classification problems.
Stéphane d'AscoliMarylou GabriéLevent SagunGiulio BiroliPublished in: NeurIPS (2021)
Keyphrases
- loss function
- data structure
- pairwise
- support vector
- learning to rank
- logistic regression
- empirical risk
- gradient boosting
- loss minimization
- reproducing kernel hilbert space
- convex loss functions
- risk minimization
- stochastic gradient descent
- hinge loss
- regularization term
- bayes rule
- semi supervised
- cost function
- machine learning