{"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/measuring-neural-net-robustness-with","title":"Measuring Neural Net Robustness with Constraints","arxiv_id":"1605.07262","date":"2016-05-24","proceeding":"NeurIPS 2016 12","authors":["Osbert Bastani","Yani Ioannou","Leonidas Lampropoulos","Dimitrios Vytiniotis","Aditya Nori","Antonio Criminisi"],"abstract":"Despite having high accuracy, neural nets have been shown to be susceptible\nto adversarial examples, where a small perturbation to an input can cause it to\nbecome mislabeled. We propose metrics for measuring the robustness of a neural\nnet and devise a novel algorithm for approximating these metrics based on an\nencoding of robustness as a linear program. We show how our metrics can be used\nto evaluate the robustness of deep neural nets with experiments on the MNIST\nand CIFAR-10 datasets. Our algorithm generates more informative estimates of\nrobustness metrics compared to estimates based on existing algorithms.\nFurthermore, we show how existing approaches to improving robustness \"overfit\"\nto adversarial examples generated using a specific algorithm. Finally, we show\nthat our techniques can be used to additionally improve neural net robustness\nboth according to the metrics that we propose, but also according to previously\nproposed metrics.","url_abs":"http://arxiv.org/abs/1605.07262v2","url_pdf":"http://arxiv.org/pdf/1605.07262v2.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":"measuring-neural-net-robustness-with","repo_url":"https://github.com/Microsoft/NeuralNetworkAnalysis","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1605.07262","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}