DPlis: Boosting Utility of Differentially Private Deep Learning via Randomized Smoothing.
Wenxiao WangTianhao WangLun WangNanqing LuoPan ZhouDawn SongRuoxi JiaPublished in: CoRR (2021)
Keyphrases
- deep learning
- differentially private
- differential privacy
- unsupervised feature learning
- machine learning
- unsupervised learning
- privacy preserving
- privacy preservation
- mental models
- data sharing
- data privacy
- weakly supervised
- learning algorithm
- personal information
- active learning
- object recognition
- feature selection
- search engine