Conference paperProtecting neural networks with hierarchical random switching: Towards better robustness-accuracy trade-off for stochastic defenses
Conference paperTrimming the 1 regularizer: Statistical analysis, optimization, and applications to deep learning