Privacy Preservation Machine Learning
隱私維護機器學習

National Taiwan University  Prof. Pei-Yuan Wu

計畫主持人:國立臺灣大學 電機工程學系 吳沛遠教授

機器學習往往牽涉機敏性資料的蒐集分析。如何避免有心人士以此窺探個人隱私遂成重要課題。本計畫旨在深度學習的架構上加入隱私維護機制,以期減少大眾對深度學習在諸如生物特徵、病歷資料等機敏性資料應用上的疑慮。

Private and sensitive data are commonly being collected and analyzed in machine learning applications. How to preserve privacy as well as preventing abusive usage of those sensitive data has become an important issue.  This research project addresses the privacy preservation issues in machine learning and deep learning architectures, as well as the attacks that are plausible given machine learning methodologies.