Tight End-to-End Accounting for Differentially Private Synthetic Data

arXiv:2610.17249 · cs.CR, cs.LG, stat.ML · Submitted 2026-10-12 · Read on arXiv

cs.CR, cs.LG, stat.ML

Submitted: 2026-10-12

Updated: 2026-10-12

Code: https://github.com/google/dpsynth

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Differentially Private — 1× in this paper · explained in 3 episodes

Sources

• Privacy and Statistical Risk: Formalisms and Minimax Bounds
• Concentrated Differential Privacy
• Privacy Filters are Captured by Residues: A Characterization of Free Natural Filters and the Cost of Adaptivity
• Fully Adaptive Composition for Gaussian Differential Privacy
• $f$-Differential Privacy Filters: Validity and Approximate Solutions

Related papers

• SoK: AI-Augmented Binary Reversing
• Relaxed Sender Anonymity for CBDC Interbank Settlement: A Zero-Knowledge Approach on Permissioned EVM
• Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages
• Efficient Fuzzy PSI under One-Sided Assumptions
• Sealing the Audit-Runtime Gap for LLM Skills
• Token Composition: A Graph Based on EVM Logs