A Multiclass Boosting Framework for Achieving Fast and Provable Adversarial Robustness.
Jacob D. AbernethyPranjal AwasthiSatyen KalePublished in: CoRR (2021)
Keyphrases
- multi class
- boosting framework
- boosting algorithms
- loss function
- pairwise
- support vector machine
- binary classification
- multi class classification
- cost sensitive
- feature selection
- base classifiers
- object detection
- semi supervised learning
- improve the classification accuracy
- pairwise constraints
- semi supervised
- data sets
- binary classifiers
- error correcting output codes
- distance function
- multi label classification
- text classification
- learning algorithm