Hybrid Defense for Deep Neural Networks: An Integration of Detecting and Cleaning Adversarial Perturbations.
Weiqi FanGuangling SunYuying SuZhi LiuXiaofeng LuPublished in: ICME Workshops (2019)
Keyphrases
- neural network
- pattern recognition
- artificial neural networks
- network security
- data fusion
- hybrid intelligent
- hybrid models
- automatic detection
- back propagation
- multi agent
- intrusion detection
- hybrid approaches
- self organizing maps
- recurrent neural networks
- learning rules
- case study
- machine learning
- genetic algorithm
- hybrid intelligent systems
- deep learning
- training process
- fuzzy systems
- evaluation function
- feed forward
- fault diagnosis
- data integration