Deep Learning with Gaussian Differential Privacy

Abstract

Deep learning models are often trained on datasets that contain sensitive information such as individuals' shopping transactions, personal contacts, and medical records. An increasingly important line of work therefore has sought to train neural networks subject to privacy constraints that are specified by differential privacy or its divergence-based relaxations. These privacy definitions, however, have weaknesses in handling certain important primitives (composition and subsampling), thereby giving loose or complicated privacy analyses of training neural networks. In this paper, we consider a recently proposed privacy definition termed \textit{$f$-differential privacy} [18] for a refined privacy analysis of training neural networks. Leveraging the appealing properties of $f$-differential privacy in handling composition and subsampling, this paper derives analytically tractable expressions for the privacy guarantees of both stochastic gradient descent and Adam used in training deep neural networks, without the need of developing sophisticated techniques as [3] did. Our results demonstrate that the $f$-differential privacy framework allows for a new privacy analysis that improves on the prior analysis~[3], which in turn suggests tuning certain parameters of neural networks for a better prediction accuracy without violating the privacy budget. These theoretically derived improvements are confirmed by our experiments in a range of tasks in image classification, text classification, and recommender systems. Python code to calculate the privacy cost for these experiments is publicly available in the \texttt{TensorFlow Privacy} library.

arXiv commentTo appear in Harvard Data Science Review

References in corpus (8)

• Adam: A Method for Stochastic Optimization
• Learning Differentially Private Recurrent Language Models
• Scalable Private Learning with PATE
• Collecting Telemetry Data Privately
• A General Approach to Adding Differential Privacy to Iterative Training Procedures
• Gaussian Differential Privacy
• Hypothesis Testing Interpretations and Renyi Differential Privacy
• Optimal Differential Privacy Composition for Exponential Mechanisms and the Cost of Adaptivity

Cited by in corpus (9)

• Privacy in Deep Learning: A Survey
• Differentially Private Learning Needs Better Features (or Much More Data)
• Understanding Gradient Clipping in Private SGD: A Geometric Perspective
• A Theoretical Perspective on Differentially Private Federated Multi-task Learning
• Efficient Privacy-Preserving Stochastic Nonconvex Optimization
• Private Stochastic Non-Convex Optimization: Adaptive Algorithms and Tighter Generalization Bounds
• Neither Private Nor Fair: Impact of Data Imbalance on Utility and Fairness in Differential Privacy
• Sharp Composition Bounds for Gaussian Differential Privacy via Edgeworth Expansion
• Data-Free Evaluation of User Contributions in Federated Learning