{"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/adversarial-manipulation-of-deep","title":"Adversarial Manipulation of Deep Representations","arxiv_id":"1511.05122","date":"2015-11-16","proceeding":null,"authors":["Sara Sabour","Yanshuai Cao","Fartash Faghri","David J. Fleet"],"abstract":"We show that the representation of an image in a deep neural network (DNN)\ncan be manipulated to mimic those of other natural images, with only minor,\nimperceptible perturbations to the original image. Previous methods for\ngenerating adversarial images focused on image perturbations designed to\nproduce erroneous class labels, while we concentrate on the internal layers of\nDNN representations. In this way our new class of adversarial images differs\nqualitatively from others. While the adversary is perceptually similar to one\nimage, its internal representation appears remarkably similar to a different\nimage, one from a different class, bearing little if any apparent similarity to\nthe input; they appear generic and consistent with the space of natural images.\nThis phenomenon raises questions about DNN representations, as well as the\nproperties of natural images themselves.","url_abs":"http://arxiv.org/abs/1511.05122v9","url_pdf":"http://arxiv.org/pdf/1511.05122v9.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":"adversarial-manipulation-of-deep","repo_url":"https://github.com/fartashf/under_convnet","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"caffe2","reach":{"status":"ok"}},{"paper_slug":"adversarial-manipulation-of-deep","repo_url":"https://github.com/rohban-lab/Salehi_submitted_2020","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1511.05122","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}