{"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/privacy-preserving-neural-representations-of","title":"Privacy-preserving Neural Representations of Text","arxiv_id":"1808.09408","date":"2018-08-28","proceeding":"EMNLP 2018 10","authors":["Maximin Coavoux","Shashi Narayan","Shay B. Cohen"],"abstract":"This article deals with adversarial attacks towards deep learning systems for\nNatural Language Processing (NLP), in the context of privacy protection. We\nstudy a specific type of attack: an attacker eavesdrops on the hidden\nrepresentations of a neural text classifier and tries to recover information\nabout the input text. Such scenario may arise in situations when the\ncomputation of a neural network is shared across multiple devices, e.g. some\nhidden representation is computed by a user's device and sent to a cloud-based\nmodel. We measure the privacy of a hidden representation by the ability of an\nattacker to predict accurately specific private information from it and\ncharacterize the tradeoff between the privacy and the utility of neural\nrepresentations. Finally, we propose several defense methods based on modified\ntraining objectives and show that they improve the privacy of neural\nrepresentations.","url_abs":"http://arxiv.org/abs/1808.09408v1","url_pdf":"http://arxiv.org/pdf/1808.09408v1.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":"privacy-preserving-neural-representations-of","repo_url":"https://github.com/mcoavoux/pnet","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"privacy-preserving","task_name":"Privacy Preserving"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1808.09408","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}