DifferentiableAbstractInterpretationforProvablyRobustNeuralNetworksMatthewMirman1TimonGehr1MartinVechev1Abstractsarialattackswhichprovidescertificatesprovingthatnoneofthetrainingexamplescouldbeadve...