Tight Certificates of Adversarial Robustness for Randomly Smoothed Classifiers.
Guang-He LeeYang YuanShiyu ChangTommi S. JaakkolaPublished in: NeurIPS (2019)
Keyphrases
- generalization error bounds
- lower bound
- upper bound
- training data
- svm classifier
- support vector
- feature selection
- classification algorithm
- supervised classification
- machine learning algorithms
- classification systems
- test set
- randomly chosen
- classifier ensemble
- ensemble learning
- classification models
- decision trees
- training samples
- feature set
- worst case
- object detection
- multi agent
- logistic regression
- class labels
- classification method
- computational efficiency
- naive bayes
- roc curve
- access control
- decision boundary
- public key
- classifier combination