Parameterizing Activation Functions for Adversarial Robustness.
Sihui DaiSaeed MahloujifarPrateek MittalPublished in: SP Workshops (2022)
Keyphrases
- activation function
- neural network
- neural architecture
- feed forward
- artificial neural networks
- complex valued
- back propagation
- feed forward neural networks
- hidden layer
- multilayer perceptron
- neural nets
- learning rate
- basis functions
- fuzzy neural network
- multi layer perceptron
- training phase
- autoregressive
- probabilistic model
- artificial intelligence