Protecting Classifiers From Attacks

Victor Gallego, Roi Naveiro, Alberto Redondo +2

In multiple domains such as malware detection, automated driving systems, or fraud detection, classification algorithms are susceptible to being attacked by malicious agents willin…