A Hybrid Training-Time and Run-Time Defense Against Adversarial Attacks in Modulation Classification.
Lu ZhangSangarapillai LambotharanGan ZhengGuisheng LiaoAmbra DemontisFabio RoliPublished in: IEEE Wirel. Commun. Lett. (2022)
Keyphrases
- training set
- supervised learning
- training process
- pattern recognition
- training phase
- support vector machine
- classification accuracy
- classification scheme
- machine learning
- support vector
- pattern classification
- automatic classification
- feature extraction
- classification method
- multi layer perceptron
- multi agent
- data sets
- benchmark datasets
- training samples
- training patterns
- image classification
- avoid overfitting
- support vector machine svm
- training examples
- feature selection
- classification algorithm
- decision trees
- machine learning methods
- classification performances
- classification models
- feature vectors
- preprocessing
- discriminative training
- sufficient training data
- classification systems
- class labels
- unsupervised learning
- online learning
- multi class
- active learning
- learning algorithm
- neural network