{"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/beyond-inferring-class-representatives-user","title":"Beyond Inferring Class Representatives: User-Level Privacy Leakage From Federated Learning","arxiv_id":"1812.00535","date":"2018-12-03","proceeding":null,"authors":["Zhibo Wang","Mengkai Song","Zhifei Zhang","Yang song","Qian Wang","Hairong Qi"],"abstract":"Federated learning, i.e., a mobile edge computing framework for deep\nlearning, is a recent advance in privacy-preserving machine learning, where the\nmodel is trained in a decentralized manner by the clients, i.e., data curators,\npreventing the server from directly accessing those private data from the\nclients. This learning mechanism significantly challenges the attack from the\nserver side. Although the state-of-the-art attacking techniques that\nincorporated the advance of Generative adversarial networks (GANs) could\nconstruct class representatives of the global data distribution among all\nclients, it is still challenging to distinguishably attack a specific client\n(i.e., user-level privacy leakage), which is a stronger privacy threat to\nprecisely recover the private data from a specific client. This paper gives the\nfirst attempt to explore user-level privacy leakage against the federated\nlearning by the attack from a malicious server. We propose a framework\nincorporating GAN with a multi-task discriminator, which simultaneously\ndiscriminates category, reality, and client identity of input samples. The\nnovel discrimination on client identity enables the generator to recover user\nspecified private data. Unlike existing works that tend to interfere the\ntraining process of the federated learning, the proposed method works\n\"invisibly\" on the server side. The experimental results demonstrate the\neffectiveness of the proposed attacking approach and the superior to the\nstate-of-the-art.","url_abs":"http://arxiv.org/abs/1812.00535v3","url_pdf":"http://arxiv.org/pdf/1812.00535v3.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":"beyond-inferring-class-representatives-user","repo_url":"https://github.com/JonasGeiping/breaching","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"edge-computing","task_name":"Edge-computing"},{"task_slug":"federated-learning","task_name":"Federated Learning"},{"task_slug":"privacy-preserving","task_name":"Privacy Preserving"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1812.00535","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}