Poison as a Cure: Detecting & Neutralizing Variable-Sized Backdoor Attacks in Deep Neural Networks.
Alvin ChanYew-Soon OngPublished in: CoRR (2019)
Keyphrases
- variable sized
- neural network
- network intrusion detection systems
- pattern recognition
- artificial neural networks
- fuzzy logic
- countermeasures
- back propagation
- self organizing maps
- traffic analysis
- early stage
- detecting malicious
- multi layer
- multilayer perceptron
- malicious attacks
- neural nets
- neural network model
- automatic detection
- genetic algorithm
- fuzzy systems
- activation function
- feed forward
- identity theft
- learning algorithm