Interpretation-Empowered Neural Cleanse for Backdoor Attacks.
Liang-Bo NingZeyu DaiJingran SuChao PanLuning WangWenqi FanQing LiPublished in: WWW (Companion Volume) (2024)
Keyphrases
- network architecture
- countermeasures
- neural network
- image interpretation
- watermarking scheme
- bio inspired
- high level
- neural computation
- security threats
- traffic analysis
- malicious attacks
- security mechanisms
- fitted q iteration
- attack detection
- artificial neural
- security risks
- biologically plausible
- biologically inspired
- learning algorithm