A General Framework For Detecting Anomalous Inputs to DNN Classifiers
Jayaram Raghuram * 1 Varun Chandrasekaran * 1 Somesh Jha 1 2 Suman Banerjee 1
Abstract side the training distribution (Nguyen et al., 2015; Szegedy
et al., 2014; Hendrycks & Gimpel, 2017; Hein et al., 2019).
Detecting anomalous inputs, such as adversarial
Such anomalous inputs can arise in real-world settings ...


雷达卡




京公网安备 11010802022788号







