PCA improves the adversarial robustness of neural networks.
István MegyeriAmmar Al-NajjarPublished in: ESANN (2022)
Keyphrases
- neural network
- principal component analysis
- principal components
- artificial neural networks
- pattern recognition
- high dimensional
- fuzzy logic
- feature extraction
- covariance matrix
- dimensionality reduction
- genetic algorithm
- robust principal component analysis
- principle component analysis
- fuzzy systems
- face recognition
- nearest neighbor
- back propagation
- expert systems
- multi layer
- learning algorithm
- appearance based object recognition
- dimension reduction methods
- principal components analysis
- activation function
- multi layer perceptron
- training process
- evaluation function
- neural network model
- dimension reduction
- independent component analysis
- face images
- computational efficiency
- fault diagnosis