Current browse context:
stat.ML
Change to browse by:
References & Citations
Statistics > Machine Learning
Title: Deep Learning with Differential Privacy
(Submitted on 1 Jul 2016 (v1), last revised 24 Oct 2016 (this version, v2))
Abstract: Machine learning techniques based on neural networks are achieving remarkable results in a wide variety of domains. Often, the training of models requires large, representative datasets, which may be crowdsourced and contain sensitive information. The models should not expose private information in these datasets. Addressing this goal, we develop new algorithmic techniques for learning and a refined analysis of privacy costs within the framework of differential privacy. Our implementation and experiments demonstrate that we can train deep neural networks with non-convex objectives, under a modest privacy budget, and at a manageable cost in software complexity, training efficiency, and model quality.
Submission history
From: Ilya Mironov [view email][v1] Fri, 1 Jul 2016 07:29:10 GMT (474kb,D)
[v2] Mon, 24 Oct 2016 11:59:40 GMT (512kb,D)
Link back to: arXiv, form interface, contact.