DI-AA: An interpretable white-box attack for fooling deep neural networks.
Yixiang WangJiqiang LiuXiaolin ChangRicardo J. RodríguezJianhua WangPublished in: Inf. Sci. (2022)
Keyphrases
- white box
- neural network
- black box
- source code
- pattern recognition
- rule extraction
- fuzzy logic
- neural nets
- source code metrics
- multilayer perceptron
- attack detection
- countermeasures
- fuzzy systems
- classification rules
- test data
- back propagation
- network architecture
- activation function
- genetic algorithm
- recurrent neural networks
- neural network model
- self organizing maps
- fault diagnosis
- deep learning
- machine learning