Tight End-to-End Accounting for Differentially Private Synthetic Data
Abstract
Differentially private (DP) pipelines often combine heterogeneous mechanisms, yet their guarantees are best analyzed using different notions of privacy (e.g., bounded range, zCDP, ADP). When composing them, we have to analyze the guarantees through a reduced common representation that loses accounting precision. This issue is particularly relevant in DP synthetic data generation, where data preprocessing, generation algorithms, and DP-SGD could appear within the same pipeline. We develop novel tools for tight end-to-end accounting of such compositions using privacy loss distributions (PLD), f-DP, and Gaussian DP (GDP). In particular, we derive PLD and GDP analyses of bounded-range mechanisms such as the Exponential mechanism. Building on these theoretical results, we introduce AIM-GDP, a GDP-native variant of AIM (a state-of-the-art DP synthetic data mechanism) that enables easy composability with complex pipelines via PLD, attack-aware risk calibration, and tight end-to-end auditing while consistently achieving better utility; and MixCal, an algorithm for calibrating adaptive mixed compositions using PLD accounting. Overall, our tools enable tighter end-to-end accounting and thus less noise at the same privacy level in DP synthetic data pipelines.
Related
- Differentially Private Synthetic Data via Generalized Moment Matching
- Polynomial-Time Differentially Private Synthetic Graph Release with Nearly-Linear Additive Error
- Differentially Private Average Treatment Effect Estimation by Propensity Score Blocking
- Communication-Efficient Differentially Private Gradient Tracking for Distributed Optimization via Local Updates
- From Privacy to Generalization: Linear Max-Information Bounds for Differentially Private Learning Algorithms
BibTeX
@misc{ganev2026tight,
title = {{Tight End-to-End Accounting for Differentially Private Synthetic Data}},
author = {Georgi Ganev and Vadym Doroshenko and Arun Ganesh and Juan Felipe Gomez and Ryan McKenna and Bogdan Kulynych},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.17249},
eprint = {2610.17249},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.17249},
}
RIS
TY - UNPB
TI - Tight End-to-End Accounting for Differentially Private Synthetic Data
AU - Georgi Ganev
AU - Vadym Doroshenko
AU - Arun Ganesh
AU - Juan Felipe Gomez
AU - Ryan McKenna
AU - Bogdan Kulynych
PY - 2026
DA - 2026/10/12
PB - arXiv
DO - 10.48550/arxiv.2610.17249
UR - https://arxiv.org/abs/2610.17249
AB - Differentially private (DP) pipelines often combine heterogeneous mechanisms, yet their guarantees are best analyzed using different notions of privacy (e.g., bounded range, zCDP, ADP). When composing them, we have to analyze the guarantees through a reduced common representation that loses accounting precision. This issue is particularly relevant in DP synthetic data generation, where data preprocessing, generation algorithms, and DP-SGD could appear within the same pipeline. We develop novel tools for tight end-to-end accounting of such compositions using privacy loss distributions (PLD), f-DP, and Gaussian DP (GDP). In particular, we derive PLD and GDP analyses of bounded-range mechanisms such as the Exponential mechanism. Building on these theoretical results, we introduce AIM-GDP, a GDP-native variant of AIM (a state-of-the-art DP synthetic data mechanism) that enables easy composability with complex pipelines via PLD, attack-aware risk calibration, and tight end-to-end auditing while consistently achieving better utility; and MixCal, an algorithm for calibrating adaptive mixed compositions using PLD accounting. Overall, our tools enable tighter end-to-end accounting and thus less noise at the same privacy level in DP synthetic data pipelines.
ER -