Aegis: Mitigating Targeted Bit-flip Attacks against Deep Neural Networks.
Jialai WangZiyuan ZhangMeiqi WangHan QiuTianwei ZhangQi LiZongpeng LiTao WeiChao ZhangPublished in: USENIX Security Symposium (2023)
Keyphrases
- neural network
- block cipher
- pattern recognition
- fuzzy logic
- neural network model
- artificial neural networks
- back propagation
- countermeasures
- risk management
- terrorist attacks
- malicious attacks
- traffic analysis
- deep learning
- watermarking scheme
- fault diagnosis
- multi layer
- neural nets
- associative memory
- cryptographic protocols
- recurrent neural networks
- feed forward
- denial of service attacks
- self organizing maps
- advanced encryption standard
- protection schemes