{"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/bolt-on-differential-privacy-for-scalable","title":"Bolt-on Differential Privacy for Scalable Stochastic Gradient Descent-based Analytics","arxiv_id":"1606.04722","date":"2016-06-15","proceeding":null,"authors":["Xi Wu","Fengan Li","Arun Kumar","Kamalika Chaudhuri","Somesh Jha","Jeffrey F. Naughton"],"abstract":"While significant progress has been made separately on analytics systems for\nscalable stochastic gradient descent (SGD) and private SGD, none of the major\nscalable analytics frameworks have incorporated differentially private SGD.\nThere are two inter-related issues for this disconnect between research and\npractice: (1) low model accuracy due to added noise to guarantee privacy, and\n(2) high development and runtime overhead of the private algorithms. This paper\ntakes a first step to remedy this disconnect and proposes a private SGD\nalgorithm to address \\emph{both} issues in an integrated manner. In contrast to\nthe white-box approach adopted by previous work, we revisit and use the\nclassical technique of {\\em output perturbation} to devise a novel \"bolt-on\"\napproach to private SGD. While our approach trivially addresses (2), it makes\n(1) even more challenging. We address this challenge by providing a novel\nanalysis of the $L_2$-sensitivity of SGD, which allows, under the same privacy\nguarantees, better convergence of SGD when only a constant number of passes can\nbe made over the data. We integrate our algorithm, as well as other\nstate-of-the-art differentially private SGD, into Bismarck, a popular scalable\nSGD-based analytics system on top of an RDBMS. Extensive experiments show that\nour algorithm can be easily integrated, incurs virtually no overhead, scales\nwell, and most importantly, yields substantially better (up to 4X) test\naccuracy than the state-of-the-art algorithms on many real datasets.","url_abs":"http://arxiv.org/abs/1606.04722v3","url_pdf":"http://arxiv.org/pdf/1606.04722v3.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":"bolt-on-differential-privacy-for-scalable","repo_url":"https://github.com/sunblaze-ucb/dpml-benchmark","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[],"methods":[{"method_slug":"sgd","method_name":"SGD"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1606.04722","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}