Provable Privacy with Non-Private Pre-Processing
Not on the map. Why?
Abstract (arXiv)
When analysing Differentially Private (DP) machine learning pipelines, the potential privacy cost of data-dependent pre-processing is frequently overlooked in privacy accounting. In this work, we propose a general framework to evaluate the additional privacy cost incurred by non-private data-dependent pre-processing algorithms. Our framework establishes upper bounds on the overall privacy guarantees by utilising two new technical notions: a variant of DP termed Smooth DP and the bounded sensitivity of the pre-processing algorithms. In addition to the generic framework, we provide explicit overall privacy guarantees for multiple data-dependent pre-processing algorithms, such as data imputation, quantization, deduplication and PCA, when used in combination with several DP algorithms. Notably, this framework is also simple to implement, allowing direct integration into existing DP pipelines.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 21 of its 43 references that are in the map. First posted on arXiv 19 Mar 2024.
Builds on
21 of 43 references.
References · 21
- 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 12,865
- 2016Deep Learning with Differential Privacycited by 2,682
- 2021Scaling Language Models: Methods, Analysis & Insights from Training Gophercited by 1,199
- 2021Deduplicating Training Data Makes Language Models Bettercited by 705
- 2018Certified Robustness to Adversarial Examples with Differential Privacycited by 462
- 2021Large Language Models Can Be Strong Differentially Private Learnerscited by 405
- 2022Deduplicating Training Data Mitigates Privacy Risks in Language Modelscited by 309
- 2020Differentially Private Learning Needs Better Features (or Much More Data)cited by 233
- 2018Privacy Amplification by Iterationcited by 97
- 2020Bypassing the Ambient Dimension: Private SGD with Gradient Subspace Identificationcited by 88
- 2021Do Not Let Privacy Overbill Utility: Gradient Embedding Perturbation for Private Learningcited by 86
- 2022Private Adaptive Optimization with Side Informationcited by 38
9 more
- 2023Privacy Side Channels in Machine Learning Systemscited by 38
- 2023Why Is Public Pretraining Necessary for Private Model Training?cited by 36
- 2020Differentially Private Simple Linear Regressioncited by 33
- 2020Investigating Membership Inference Attacks under Data Dependenciescited by 27
- 2022DP-PCA: Statistically Optimal and Differentially Private PCAcited by 24
- 2016Pufferfish Privacy Mechanisms for Correlated Datacited by 16
- 2023PILLAR: How to make semi-private learning more effectivecited by 10
- 2016Statistical Properties of Sanitized Results from Differentially Private Laplace Mechanism with Univariate Bounding Constraintscited by 4
- 2023R'enyi Pufferfish Privacy: General Additive Noise Mechanisms and Privacy Amplification by Iterationcited by 2
Built on by
Citing papers per year: 2024: 1; 2025: 1.
2 papers.