{"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/interpreting-adversarial-examples-with","title":"Interpreting Adversarial Examples with Attributes","arxiv_id":"1904.08279","date":"2019-04-17","proceeding":null,"authors":["Sadaf Gulshad","Jan Hendrik Metzen","Arnold Smeulders","Zeynep Akata"],"abstract":"Deep computer vision systems being vulnerable to imperceptible and carefully\ncrafted noise have raised questions regarding the robustness of their\ndecisions. We take a step back and approach this problem from an orthogonal\ndirection. We propose to enable black-box neural networks to justify their\nreasoning both for clean and for adversarial examples by leveraging attributes,\ni.e. visually discriminative properties of objects. We rank attributes based on\ntheir class relevance, i.e. how the classification decision changes when the\ninput is visually slightly perturbed, as well as image relevance, i.e. how well\nthe attributes can be localized on both clean and perturbed images. We present\ncomprehensive experiments for attribute prediction, adversarial example\ngeneration, adversarially robust learning, and their qualitative and\nquantitative analysis using predicted attributes on three benchmark datasets.","url_abs":"http://arxiv.org/abs/1904.08279v1","url_pdf":"http://arxiv.org/pdf/1904.08279v1.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":"interpreting-adversarial-examples-with","repo_url":"https://github.com/sadafgulshad1/Understaning-Misclassifications-by-Attributes","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"attribute","task_name":"Attribute"},{"task_slug":"classification","task_name":"General Classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}