A General Approach to Adding Differential Privacy to Iterative Training Procedures

Abstract

In this work we address the practical challenges of training machine learning models on privacy-sensitive datasets by introducing a modular approach that minimizes changes to training algorithms, provides a variety of configuration strategies for the privacy mechanism, and then isolates and simplifies the critical logic that computes the final privacy guarantees. A key challenge is that training algorithms often require estimating many different quantities (vectors) from the same set of examples --- for example, gradients of different layers in a deep learning architecture, as well as metrics and batch normalization parameters. Each of these may have different properties like dimensionality, magnitude, and tolerance to noise. By extending previous work on the Moments Accountant for the subsampled Gaussian mechanism, we can provide privacy for such heterogeneous sets of vectors, while also structuring the approach to minimize software engineering challenges.

Presented at NeurIPS 2018 workshop on Privacy Preserving Machine Learning; Companion paper to TensorFlow Privacy OSS Library

Cited by in corpus (14)

• How to DP-fy ML: A Practical Guide to Machine Learning with Differential Privacy
• Do Gradient Inversion Attacks Make Federated Learning Unsafe?
• Adversarial Attacks Against Deep Generative Models on Data: A Survey
• Differentially Private Learning with Adaptive Clipping
• Differential Privacy Has Disparate Impact on Model Accuracy
• DataLens: Scalable Privacy Preserving Training via Gradient Compression and Aggregation
• Privacy Accounting and Quality Control in the Sage Differentially Private ML Platform
• Deep Learning with Gaussian Differential Privacy
• Learning discrete distributions: user vs item-level privacy
• Differentially Private Multivariate Time Series Forecasting of Aggregated Human Mobility With Deep Learning: Input or Gradient Perturbation?
• Differentially Private Federated Learning with Laplacian Smoothing
• Fairness-aware Differentially Private Collaborative Filtering
• Machine Learning Systems for Intelligent Services in the IoT: A Survey
• Exploring the Unfairness of DP-SGD Across Settings