Distillation as a Defense to Adversarial Perturbations Against Deep Neural Networks.
Nicolas PapernotPatrick D. McDanielXi WuSomesh JhaAnanthram SwamiPublished in: IEEE Symposium on Security and Privacy (2016)
Keyphrases
- neural network
- pattern recognition
- multi layer
- artificial neural networks
- back propagation
- intrusion detection
- training process
- multi agent
- self organizing maps
- competitive learning
- data mining
- databases
- rule extraction
- neural nets
- deep learning
- multilayer perceptron
- hopfield neural network
- real time
- neural network model
- feed forward
- fault diagnosis
- recurrent neural networks
- network security
- association rules
- expert systems
- fuzzy systems
- training data
- activation function
- information systems
- search engine
- learning algorithm
- neuro fuzzy systems