Abstract: Notwithstanding the tremendous success of deep neural networks in a range of realms, previous studies have shown that these learning models are exposed to an inherent hazard called ...