Interpolated Adversarial Training: Achieving Robust Neural Networks Without Sacrificing Too Much Accuracy.
Alex LambVikas VermaJuho KannalaYoshua BengioPublished in: AISec@CCS (2019)
Keyphrases
- neural network
- training algorithm
- training process
- high accuracy
- feed forward neural networks
- highly accurate
- pattern recognition
- multi layer perceptron
- estimation error
- feedforward neural networks
- radial basis function network
- back propagation
- backpropagation algorithm
- artificial neural networks
- neural network model
- multi agent
- supervised learning
- computational efficiency
- multilayer perceptron
- training speed
- computational cost
- multi layer
- active learning
- linear interpolation
- error back propagation
- training phase
- fuzzy systems
- neural nets
- self organizing maps
- test set
- fuzzy logic
- classification accuracy
- machine learning