← 研踪 PaperTrailarXiv:2602.06756 · v3
f-Differential Privacy Filters: Validity and Approximate Solutions
首次提交 2026-02-06 23:04
版本更新 2026-11-07 20:35
雷达观测 2026-11-10 11:14
摘要
Accounting for privacy loss under fully adaptive composition---where mechanism choice and privacy parameters may depend on the history of prior outputs---is a central challenge in differential privacy (DP). Here, privacy filters are stopping rules ensuring a prescribed global budget is not exceeded. A leading candidate for optimal filter design is -DP, which characterizes the full extent of adversarial hypothesis testing and recovers -DP through piece-wise linear trade-off functions, while enabling tight -DP accounting in standard compositions via tensor products. Yet whether such filters can be correctly defined under -DP remains unclear. We show that the natural -DP filter---tracking path-wise accumulating tensor products and stopping when the prescribed curve is crossed---is fundamentally invalid, precluding the direct use of standard efficient numerical Fast-Fourier-Transform accounting in the fully adaptive setting. We characterize this failure, establishing necessary and sufficient conditions for the natural filter's validity. Furthermore, we prove a fully adaptive central limit theorem for -DP, establishing Gaussian convergence of cumulative privacy losses under full adaptivity. As a demonstration, we construct a closed-form approximate GDP filter for subsampled Gaussian mechanisms that provably outperforms RDP-based accounting in asymptotic regimes ( and ) without tracking the full trade-off function, demonstrating that the slack in RDP is not intrinsic to adaptive composition---though CLT-based approximations are known to be optimistic at realistic subsampling rates, a limitation that remains an open challenge.
作者备注:NeurIPS 2026. 54 pages, 12 figures
观测到的版本
相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。
v3 · 2026-11-07 20:35 · 最新观测版本
f-Differential Privacy Filters: Validity and Approximate Solutions
Accounting for privacy loss under fully adaptive composition---where mechanism choice and privacy parameters may depend on the history of prior outputs---is a central challenge in differential privacy (DP). Here, privacy filters are stopping rules ensuring a prescribed global budget is not exceeded. A leading candidate for optimal filter design is -DP, which characterizes the full extent of adversarial hypothesis testing and recovers -DP through piece-wise linear trade-off functions, while enabling tight -DP accounting in standard compositions via tensor products. Yet whether such filters can be correctly defined under -DP remains unclear. We show that the natural -DP filter---tracking path-wise accumulating tensor products and stopping when the prescribed curve is crossed---is fundamentally invalid, precluding the direct use of standard efficient numerical Fast-Fourier-Transform accounting in the fully adaptive setting. We characterize this failure, establishing necessary and sufficient conditions for the natural filter's validity. Furthermore, we prove a fully adaptive central limit theorem for -DP, establishing Gaussian convergence of cumulative privacy losses under full adaptivity. As a demonstration, we construct a closed-form approximate GDP filter for subsampled Gaussian mechanisms that provably outperforms RDP-based accounting in asymptotic regimes ( and ) without tracking the full trade-off function, demonstrating that the slack in RDP is not intrinsic to adaptive composition---though CLT-based approximations are known to be optimistic at realistic subsampling rates, a limitation that remains an open challenge.
查看此版本解读 →查看此版本原文 ↗