Adversarial examples represent a major security threat for emerging technologies that use machine learning models to make important decisions. Adversarial examples are typically crafted by performing gradient descent on the target model, but recently, researchers have started to look at methods to generate adversarial examples without requiring access to the target model at decision time. This thesis investigates the place that generative attacks have in the machine learning security threat landscape and improvements that can be made on existing attacks. By evaluating attacks and defenses on image classiﬁcation problems, we found that generative attackers are most relevant in black box settings where query access is given to the attacker before test time. We evaluate generative methods in these settings as standalone attacks and we give examples of how generative attacks can reduce the number of queries or amount of time required to produce a successful adversarial example by narrowing the search space of existing optimization algorithms.
|Advisor:||Fontaine, Fred L.|
|Commitee:||Shoop, Barry L.|
|School:||The Cooper Union for the Advancement of Science and Art|
|School Location:||United States -- New York|
|Source:||MAI 82/4(E), Masters Abstracts International|
|Subjects:||Artificial intelligence, Electrical engineering, Computer science|
|Keywords:||Adversarial Examples, Computer Security, Deep Learning, Generative Attacks, Machine Learning|
Copyright in each Dissertation and Thesis is retained by the author. All Rights Reserved
The supplemental file or files you are about to download were provided to ProQuest by the author as part of a
dissertation or thesis. The supplemental files are provided "AS IS" without warranty. ProQuest is not responsible for the
content, format or impact on the supplemental file(s) on our system. in some cases, the file type may be unknown or
may be a .exe file. We recommend caution as you open such files.
Copyright of the original materials contained in the supplemental file is retained by the author and your access to the
supplemental files is subject to the ProQuest Terms and Conditions of use.
Depending on the size of the file(s) you are downloading, the system may take some time to download them. Please be