Near-optimal Evasion of Randomized Convex-inducing Classifiers in Adversarial Environments.
Pooria MadaniNatalija VlajicPublished in: ARES (2019)
Keyphrases
- feature selection
- training data
- decision trees
- linear classifiers
- support vector
- machine learning algorithms
- convex optimization
- training set
- class labels
- dynamic environments
- test set
- majority vote
- training samples
- feature set
- multi agent
- convex functions
- decision boundary
- extracted features
- classification rate
- piecewise linear
- real world
- classifier combination
- convex relaxation
- weak classifiers
- countermeasures
- feature extraction
- ensemble learning
- classification models
- data sets
- semi supervised learning
- naive bayes