Abstract: As deep neural networks (DNNs) are becoming vital to numerous safety-critical applications, ensuring their fault reliability is crucial. This paper presents a hybrid framework that combines ...