Enhancing Transformation-Based Defenses Against Adversarial Attacks with a Distribution Classifier.
Connie KouHwee Kuan LeeEe-Chien ChangTeck Khim NgPublished in: ICLR (2020)
Keyphrases
- spam filters
- denial of service attacks
- countermeasures
- classification scheme
- decision trees
- training set
- training samples
- class labels
- feature selection
- training data
- multi agent
- feature set
- classification rate
- classification algorithm
- classification method
- denial of service
- classification process
- dos attacks
- training examples
- feature space
- support vector
- label noise
- network layer
- mixed data
- class membership
- class conditional
- network security
- multiple classifiers
- classifier ensemble
- security mechanisms
- watermarking scheme
- data mining
- test data
- data distribution
- learning algorithm