{"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/logan-membership-inference-attacks-against","title":"LOGAN: Membership Inference Attacks Against Generative Models","arxiv_id":"1705.07663","date":"2017-05-22","proceeding":null,"authors":["Jamie Hayes","Luca Melis","George Danezis","Emiliano De Cristofaro"],"abstract":"Generative models estimate the underlying distribution of a dataset to\ngenerate realistic samples according to that distribution. In this paper, we\npresent the first membership inference attacks against generative models: given\na data point, the adversary determines whether or not it was used to train the\nmodel. Our attacks leverage Generative Adversarial Networks (GANs), which\ncombine a discriminative and a generative model, to detect overfitting and\nrecognize inputs that were part of training datasets, using the discriminator's\ncapacity to learn statistical differences in distributions.\n  We present attacks based on both white-box and black-box access to the target\nmodel, against several state-of-the-art generative models, over datasets of\ncomplex representations of faces (LFW), objects (CIFAR-10), and medical images\n(Diabetic Retinopathy). We also discuss the sensitivity of the attacks to\ndifferent training parameters, and their robustness against mitigation\nstrategies, finding that defenses are either ineffective or lead to\nsignificantly worse performances of the generative models in terms of training\nstability and/or sample quality.","url_abs":"http://arxiv.org/abs/1705.07663v4","url_pdf":"http://arxiv.org/pdf/1705.07663v4.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":"logan-membership-inference-attacks-against","repo_url":"https://github.com/jhayes14/gen_mem_inf","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1705.07663","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}