Concept-based Adversarial Attacks: Tricking Humans and Classifiers Alike.
Johannes SchneiderGiovanni ApruzzesePublished in: SP Workshops (2022)
Keyphrases
- support vector
- training data
- decision trees
- countermeasures
- machine learning algorithms
- training set
- supervised classification
- multiple classifiers
- classifier combination
- text categorization
- feature selection
- linear classifiers
- test set
- classification systems
- security threats
- classification algorithm
- human behavior
- human users
- denial of service attacks
- dos attacks
- svm classifier
- class labels
- naive bayes
- feature set
- chosen plaintext
- malicious users
- security mechanisms
- binary classifiers
- weak classifiers
- classification rate
- classification models
- machine learning methods
- logistic regression
- semantic information
- semi supervised
- multi agent