Dimensionality Reduction as a Defense against Evasion Attacks on Machine Learning Classifiers.
Arjun Nitin BhagojiDaniel CullinaPrateek MittalPublished in: CoRR (2017)
Keyphrases
- dimensionality reduction
- machine learning
- countermeasures
- feature selection
- machine learning algorithms
- machine learning methods
- decision trees
- machine learning approaches
- semi supervised dimensionality reduction
- pattern recognition
- defense mechanisms
- supervised classification
- ddos attacks
- support vector
- dimensionality reduction methods
- data representation
- information security
- support vector machine
- learning tasks
- learning classifier systems
- learning algorithm
- principal components
- svm classifier
- computer vision
- feature extraction
- intrusion detection
- information extraction
- principal component analysis
- low dimensional
- inductive learning
- training data
- high dimensionality
- naive bayes
- high dimensional data
- pattern recognition and machine learning
- structure preserving
- text classification
- training set
- high dimensional
- network security
- active learning
- learning problems
- multiple classifier systems
- semi supervised
- supervised learning
- training samples
- feature set
- class labels
- classification models
- manifold learning
- classifier combination
- feature subset
- natural language processing
- roc analysis
- classification algorithm
- linear discriminant analysis
- text categorization