The Space of Transferable Adversarial Examples
Florian Tramèr Affiliation: Stanford University Nicolas Papernot Affiliation: Pennsylvania State University Ian Goodfellow Affiliation: Google Brain Dan Boneh Affiliation: Stanford University Patrick McDaniel Affiliation: Pennsylvania State University
Abstract
Adversarial examples are maliciously perturbed inputs designed to mislead machine learning (ML) models at test-time. They often transfer: the same adversarial example fools more than one model.
原文 arXiv:1704.03453;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1704.03453v2