{"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/cost-sensitive-robustness-against-adversarial","title":"Cost-Sensitive Robustness against Adversarial Examples","arxiv_id":"1810.09225","date":"2018-10-22","proceeding":"ICLR 2019 5","authors":["Xiao Zhang","David Evans"],"abstract":"Several recent works have developed methods for training classifiers that are\ncertifiably robust against norm-bounded adversarial perturbations. These\nmethods assume that all the adversarial transformations are equally important,\nwhich is seldom the case in real-world applications. We advocate for\ncost-sensitive robustness as the criteria for measuring the classifier's\nperformance for tasks where some adversarial transformation are more important\nthan others. We encode the potential harm of each adversarial transformation in\na cost matrix, and propose a general objective function to adapt the robust\ntraining method of Wong & Kolter (2018) to optimize for cost-sensitive\nrobustness. Our experiments on simple MNIST and CIFAR10 models with a variety\nof cost matrices show that the proposed approach can produce models with\nsubstantially reduced cost-sensitive robust error, while maintaining\nclassification accuracy.","url_abs":"http://arxiv.org/abs/1810.09225v2","url_pdf":"http://arxiv.org/pdf/1810.09225v2.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":"cost-sensitive-robustness-against-adversarial","repo_url":"https://github.com/xiaozhanguva/Cost-Sensitive-Robustness","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"classification","task_name":"General Classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1810.09225","atlas_url":"https://app.syntology.ai/?focus=1810.09225","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}