STRIP: a defence against trojan attacks on deep neural networks.
Yansong GaoChange XuDerui WangShiping ChenDamith Chinthana RanasingheSurya NepalPublished in: ACSAC (2019)
Keyphrases
- neural network
- pattern recognition
- genetic algorithm
- artificial neural networks
- fuzzy logic
- back propagation
- countermeasures
- malicious users
- security threats
- multi layer
- neural network model
- recurrent neural networks
- network analysis
- fuzzy systems
- deep learning
- social networks
- multilayer perceptron
- training process
- self organizing maps
- watermarking algorithm
- rule extraction
- ddos attacks
- fault diagnosis
- traffic analysis