Adversarial robustness via stochastic regularization of neural activation sensitivity.
Gil FidelRon BittonZiv KatzirAsaf ShabtaiPublished in: CoRR (2020)
Keyphrases
- neural network
- lateral inhibition
- network architecture
- information processing
- monte carlo
- computational efficiency
- sensitivity analysis
- high sensitivity
- bayesian networks
- biologically inspired
- stochastic model
- neural model
- stochastic optimization
- stochastic programming
- biologically plausible
- regularization methods
- connectionist models
- stability of feature selection
- prior information