{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/bridging-machine-learning-and-cryptography-in","title":"Bridging machine learning and cryptography in defence against adversarial attacks","arxiv_id":"1809.01715","date":"2018-09-05","proceeding":null,"authors":["Olga Taran","Shideh Rezaeifar","Slava Voloshynovskiy"],"abstract":"In the last decade, deep learning algorithms have become very popular thanks\nto the achieved performance in many machine learning and computer vision tasks.\nHowever, most of the deep learning architectures are vulnerable to so called\nadversarial examples. This questions the security of deep neural networks (DNN)\nfor many security- and trust-sensitive domains. The majority of the proposed\nexisting adversarial attacks are based on the differentiability of the DNN cost\nfunction.Defence strategies are mostly based on machine learning and signal\nprocessing principles that either try to detect-reject or filter out the\nadversarial perturbations and completely neglect the classical cryptographic\ncomponent in the defence. In this work, we propose a new defence mechanism\nbased on the second Kerckhoffs's cryptographic principle which states that the\ndefence and classification algorithm are supposed to be known, but not the key.\nTo be compliant with the assumption that the attacker does not have access to\nthe secret key, we will primarily focus on a gray-box scenario and do not\naddress a white-box one. More particularly, we assume that the attacker does\nnot have direct access to the secret block, but (a) he completely knows the\nsystem architecture, (b) he has access to the data used for training and\ntesting and (c) he can observe the output of the classifier for each given\ninput. We show empirically that our system is efficient against most famous\nstate-of-the-art attacks in black-box and gray-box scenarios.","url_abs":"http://arxiv.org/abs/1809.01715v1","url_pdf":"http://arxiv.org/pdf/1809.01715v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"bridging-machine-learning-and-cryptography-in","repo_url":"https://github.com/GrandGarcon/Complete_Cybersecurity_Path","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"bridging-machine-learning-and-cryptography-in","repo_url":"https://github.com/SimSteve/Project","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"bridging-machine-learning-and-cryptography-in","repo_url":"https://github.com/nikste/crypto-defense","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"machine-learning","task_name":"BIG-bench Machine Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1809.01715","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}