Training Confidence-calibrated Classifiers for Detecting Out-of-Distribution Samples.
Kimin LeeHonglak LeeKibok LeeJinwoo ShinPublished in: ICLR (Poster) (2018)
Keyphrases
- training samples
- training set
- training examples
- test set
- training data
- training and testing data
- number of training samples
- training and test data
- sample set
- test sample
- optimum path forest
- test data
- support vector
- data samples
- confidence scores
- classifier training
- training set size
- supervised learning
- feature space
- class conditional
- discriminative classifiers
- avoid overfitting
- class distribution
- labeled training data
- training dataset
- small sample
- classification algorithm
- svm classifier
- support vector machine
- trained classifiers
- learning algorithm
- confidence measures
- sample selection
- boosted classifiers
- text classifiers
- decision trees
- probability distribution
- machine learning algorithms
- linear classifiers
- weak classifiers
- training process
- density estimates
- hyperplane
- probability density function
- object detectors
- naive bayes
- labeled data
- feature set
- classification performances
- decision boundary
- classification accuracy
- confidence measure
- feature selection
- neural network