Construction of non-convex polynomial loss functions for training a binary classifier with quantum annealing.
Ryan BabbushVasil S. DenchevNan DingSergei IsakovHartmut NevenPublished in: CoRR (2014)
Keyphrases
- loss function
- empirical risk
- risk minimization
- hinge loss
- support vector
- partially labeled data
- pairwise
- training process
- training set
- binary classifiers
- convex loss functions
- multi class
- training samples
- training examples
- loss minimization
- logistic regression
- bayes rule
- squared error
- learning to rank
- kernel classifiers
- simulated annealing
- support vector machine
- bregman divergences
- linear classifiers
- error correcting output codes
- decision function
- pairwise constraints
- training data
- linear svm
- reproducing kernel hilbert space
- decision trees
- training dataset
- feature set
- line search
- boosting algorithms
- training algorithm
- svm classifier
- semi supervised