Adversarial attacks and defenses using feature-space stochasticity.
Jumpei UkitaKenichi OhkiPublished in: Neural Networks (2023)
Keyphrases
- feature space
- denial of service attacks
- denial of service
- dos attacks
- input space
- spam filters
- principal component analysis
- countermeasures
- dimensionality reduction
- high dimensional
- feature selection
- network security
- training set
- high dimensionality
- feature vectors
- training samples
- malicious attacks
- mean shift
- classification accuracy
- traffic analysis
- feature extraction
- lightweight
- support vector machine
- watermarking scheme
- watermarking algorithm
- stochastic model
- kernel function
- image representation
- image retrieval
- dissimilarity measure
- spam filtering
- security protocols
- ddos attacks
- high dimensional feature space
- security threats
- input data
- low dimensional
- wireless sensor networks
- decision trees
- terrorist attacks
- attack detection
- machine learning