Optimizing Canaries for Privacy Auditing with Metagradient Descent

• URL: http://arxiv.org/abs/2507.15836v1
• Date: Mon, 21 Jul 2025 17:47:33 GMT
• Title: Optimizing Canaries for Privacy Auditing with Metagradient Descent
• Authors: Matteo Boglioni, Terrance Liu, Andrew Ilyas, Zhiwei Steven Wu,
• Abstract summary: We study black-box privacy auditing, where the goal is to lower bound the privacy parameter of a differentially private learning algorithm.<n>Our main contribution is a method for optimizing the auditor's canary set to improve privacy auditing.<n>Our empirical evaluation demonstrates that by using such optimized canaries, we can improve empirical lower bounds for differentially private image classification models by over 2x in certain instances.
• Score: 32.69637681449977
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: In this work we study black-box privacy auditing, where the goal is to lower bound the privacy parameter of a differentially private learning algorithm using only the algorithm's outputs (i.e., final trained model). For DP-SGD (the most successful method for training differentially private deep learning models), the canonical approach auditing uses membership inference-an auditor comes with a small set of special "canary" examples, inserts a random subset of them into the training set, and then tries to discern which of their canaries were included in the training set (typically via a membership inference attack). The auditor's success rate then provides a lower bound on the privacy parameters of the learning algorithm. Our main contribution is a method for optimizing the auditor's canary set to improve privacy auditing, leveraging recent work on metagradient optimization. Our empirical evaluation demonstrates that by using such optimized canaries, we can improve empirical lower bounds for differentially private image classification models by over 2x in certain instances. Furthermore, we demonstrate that our method is transferable and efficient: canaries optimized for non-private SGD with a small model architecture remain effective when auditing larger models trained with DP-SGD.

Related papers

• Enhancing One-run Privacy Auditing with Quantile Regression-Based Membership Inference [22.843200081364873]
  Differential privacy (DP) auditing aims to provide empirical lower bounds on the privacy guarantees of DP mechanisms like DP-SGD.<n>Recent work introduces one-run auditing approaches that effectively audit DP-SGD in white-box settings while still being computationally efficient.<n>In this work, we study how incorporating approaches for stronger membership inference attacks (MIA) can improve one-run auditing in the black-box setting.
  arXiv  Detail & Related papers  (2025-06-18T11:03:39Z)
• Federated Learning with Differential Privacy: An Utility-Enhanced Approach [12.614480013684759]
  Federated learning has emerged as an attractive approach to protect data privacy by eliminating the need for sharing clients' data.<n>Recent studies have shown that federated learning alone does not guarantee privacy, as private data may still be inferred from the uploaded parameters to the central server.<n>We present a modification to these vanilla differentially private algorithms based on a Haar wavelet transformation step and a novel noise injection scheme that significantly lowers the bound of the noise variance.
  arXiv  Detail & Related papers  (2025-03-27T04:48:29Z)
• DiSK: Differentially Private Optimizer with Simplified Kalman Filter for Noise Reduction [57.83978915843095]
  This paper introduces DiSK, a novel framework designed to significantly enhance the performance of differentially private gradients.<n>To ensure practicality for large-scale training, we simplify the Kalman filtering process, minimizing its memory and computational demands.
  arXiv  Detail & Related papers  (2024-10-04T19:30:39Z)
• Too Good to be True? Turn Any Model Differentially Private With DP-Weights [0.0]
  We introduce a groundbreaking approach that applies differential privacy noise to the model's weights after training.<n>We offer a comprehensive mathematical proof for this novel approach's privacy bounds.<n>We empirically evaluate its effectiveness using membership inference attacks and performance evaluations.
  arXiv  Detail & Related papers  (2024-06-27T19:58:11Z)
• Tight Auditing of Differentially Private Machine Learning [77.38590306275877]
  For private machine learning, existing auditing mechanisms are tight. They only give tight estimates under implausible worst-case assumptions. We design an improved auditing scheme that yields tight privacy estimates for natural (not adversarially crafted) datasets.
  arXiv  Detail & Related papers  (2023-02-15T21:40:33Z)
• One-shot Empirical Privacy Estimation for Federated Learning [43.317478030880956]
  "One-shot" approach allows efficient auditing or estimation of the privacy loss of a model during the same, single training run used to fit model parameters. We show that our method provides provably correct estimates for the privacy loss under the Gaussian mechanism.
  arXiv  Detail & Related papers  (2023-02-06T19:58:28Z)
• On the utility and protection of optimization with differential privacy and classic regularization techniques [9.413131350284083]
  We study the effectiveness of the differentially-private descent (DP-SGD) algorithm against standard optimization practices with regularization techniques. We discuss differential privacy's flaws and limits and empirically demonstrate the often superior privacy-preserving properties of dropout and l2-regularization.
  arXiv  Detail & Related papers  (2022-09-07T14:10:21Z)
• Individual Privacy Accounting for Differentially Private Stochastic Gradient Descent [69.14164921515949]
  We characterize privacy guarantees for individual examples when releasing models trained by DP-SGD. We find that most examples enjoy stronger privacy guarantees than the worst-case bound. This implies groups that are underserved in terms of model utility simultaneously experience weaker privacy guarantees.
  arXiv  Detail & Related papers  (2022-06-06T13:49:37Z)
• Large Scale Transfer Learning for Differentially Private Image Classification [51.10365553035979]
  Differential Privacy (DP) provides a formal framework for training machine learning models with individual example level privacy. Private training using DP-SGD protects against leakage by injecting noise into individual example gradients. While this result is quite appealing, the computational cost of training large-scale models with DP-SGD is substantially higher than non-private training.
  arXiv  Detail & Related papers  (2022-05-06T01:22:20Z)
• Large Language Models Can Be Strong Differentially Private Learners [70.0317718115406]
  Differentially Private (DP) learning has seen limited success for building large deep learning models of text. We show that this performance drop can be mitigated with the use of large pretrained models. We propose a memory saving technique that allows clipping in DP-SGD to run without instantiating per-example gradients.
  arXiv  Detail & Related papers  (2021-10-12T01:45:27Z)
• Tempered Sigmoid Activations for Deep Learning with Differential Privacy [33.574715000662316]
  We show that the choice of activation function is central to bounding the sensitivity of privacy-preserving deep learning. We achieve new state-of-the-art accuracy on MNIST, FashionMNIST, and CIFAR10 without any modification of the learning procedure fundamentals.
  arXiv  Detail & Related papers  (2020-07-28T13:19:45Z)
• User-Level Privacy-Preserving Federated Learning: Analysis and Performance Optimization [77.43075255745389]
  Federated learning (FL) is capable of preserving private data from mobile terminals (MTs) while training the data into useful models. From a viewpoint of information theory, it is still possible for a curious server to infer private information from the shared models uploaded by MTs. We propose a user-level differential privacy (UDP) algorithm by adding artificial noise to the shared models before uploading them to servers.
  arXiv  Detail & Related papers  (2020-02-29T10:13:39Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.