Increasing-Margin Adversarial (IMA) Training to Improve Adversarial Robustness of Neural Networks.
Linhai MaLiang LiangPublished in: CoRR (2020)
Keyphrases
- neural network
- training algorithm
- artificial neural networks
- training set
- multi agent
- training process
- sufficient conditions
- neural network training
- backpropagation algorithm
- feed forward neural networks
- recurrent neural networks
- computational efficiency
- supervised learning
- pattern recognition
- support vector
- real time
- computational intelligence
- feed forward
- multi layer
- fuzzy systems
- semi supervised
- fuzzy logic
- multi layer perceptron
- case study
- neural network structure