.png){kind=icon}
返回
A General Framework for Adversarial Examples with Objectives
.png){kind=icon}
OA
AI
DOI：10.1145/3317611.png)
.png){kind=icon}
.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Images perturbed subtly to be misclassified by neural networks, called adversarial examples, have emerged as a technically deep challenge and an important concern for several application domains. Most research on adversarial examples takes as its only constraint that the perturbed images are similar to the originals. However, real-world application of these ideas often requires the examples to satisfy additional objectives, which are typically enforced through custom modifications of the perturbation process. In this article, we propose adversarial generative nets (AGNs), a general methodology to train a generator neural network to emit adversarial examples satisfying desired objectives. We demonstrate the ability of AGNs to accommodate a wide range of objectives, including imprecise ones difficult to model, in two application domains. In particular, we demonstrate physical adversarial examples-eyeglass frames designed to fool face recognition-with better robustness, inconspicuousness, and scalability than previous approaches, as well as a new attack to fool a handwritten-digit classifier.
Keyword:
Machine learning
neural networks
face recognition
adversarial examples
AI总结
对已上传原文的论文进行重点信息的提取，主要内容包括：简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。
期刊
A
IF:
2.8
论文数:
293
被引数:
770
机构
引用论文
Welding characteristics of aluminum, copper, nickel and aluminum alloy with alumina coating using ultrasonic complex vibration welding equipments铝、铜、镍及铝合金氧化铝涂层超声复合振动焊接特性研究
PREAI
分享收藏An experimental study on the environmental impact of hydrogen and natural gas blend burning.png)
Chemosphere
ChemosphereIF0
PREAI
分享收藏Different rates of development associated with the alcohol dehydrogenase locus in the seaweed fly, Coelopa frigidaHeredityOAAI
HeredityIF0
OAAI分享收藏
查看更多内容