On the Minimal Adversarial Perturbation for Deep Neural Networks with Provable Estimation Error.
Fabio BrauGiulio RossoliniAlessandro BiondiGiorgio C. ButtazzoPublished in: CoRR (2022)
Keyphrases
- estimation error
- neural network
- estimation accuracy
- pattern recognition
- average error
- error rate
- approximation error
- neural network model
- genetic algorithm
- artificial neural networks
- standard deviation
- fault diagnosis
- covariance matrix
- training process
- multi agent
- back propagation
- evolutionary algorithm
- fuzzy logic
- self organizing maps
- objective function
- activation function
- multilayer perceptron
- neural nets
- multi layer
- learning algorithm
- multiscale
- feed forward
- lower bound
- prediction accuracy
- principal component analysis