Fetching the paper…

MULDEF: Multi-model-based Defense Against Adversarial Examples for Neural Networks · Around