Improving Classifier Confidence using Lossy Label-Invariant Transformations.
Sooyong JangInsup LeeJames WeimerPublished in: AISTATS (2021)
Keyphrases
- class labels
- label noise
- preserving transformations
- high confidence
- quasi invariant
- multi label classification
- feature selection
- confidence estimates
- similarity transformations
- geometric transformations
- classification process
- data compression
- support vector
- image transformations
- linear classifiers
- learning algorithm
- multilabel classification
- affine invariant
- decision trees
- multi label
- classification method
- classification algorithm
- feature space
- semantic labels
- training set
- confidence values
- invariant representations
- class membership
- confidence measure
- training data
- classification rate
- classification scheme
- ensemble learning
- machine learning
- wavelet transform
- confidence scores
- confidence level
- training examples
- image coding
- invariant features