Titlebook: Machine Learning and Knowledge Discovery in Databases. Research Track; European Conference, Albert Bifet,Jesse Davis,Indrė Žliobaitė Confer

有节制

A Theoretically Grounded Extension of Universal Attacks from the Attacker’s ViewpointSpecifically, we embrace the attacker’s perspective and introduce a theoretical bound quantifying how much the universal perturbations are able to fool a given model on unseen examples. An extension to assert the transferability of universal attacks is also provided. To learn such perturbations, we

马赛克

Linear Modeling of the Adversarial Noise Spaceinguish specific attacks adapted to each example from universal ones, referred to as example-agnostic. Even though specific adversarial attacks are efficient on their target DNN classifier to attack, they struggle to transfer to others. Conversely, universal adversarial attacks suffer from lower att

偏见

交响乐