A Generalized Meta-loss Function for Distillation Based Learning Using Privileged Information for Classification and Regression.
Amina AsifMuhammad DawoodFayyaz ul Amir Afsar MinhasPublished in: ICANN (3) (2021)
Keyphrases
- loss function
- bayes rule
- privileged information
- histogram intersection kernel
- support vector
- supervised learning
- learning algorithm
- gradient boosting
- classification accuracy
- boosting framework
- learning tasks
- feature vectors
- linear regression
- logistic regression
- risk minimization
- stochastic gradient descent
- base learners
- learning problems
- regression model
- support vector machine svm
- active learning
- training data
- feature extraction
- feature selection