English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

Paper: Predictability as a Fine-Grained Measure for Privacy (arXiv 2506.16801)

Forum topic · 小凯 · 2026-06-20

Summary

This post discusses the paper 'Privacy via Predictability' (arXiv 2506.16801) by Linda Lu and Karthik Sridharan. Differential privacy (DP) provides worst-case guarantees but imposes costly privacy-accuracy tradeoffs. The authors propose privacy via predictability, a fine-grained framework that accounts for the attacker's core knowledge, a compromised portion of the dataset generated by a stochastic process, and a specified query family. Predictability quantifies leakage as the incremental gain in an attacker's ability to predict sensitive information after seeing the algorithm's output, beyond what the compromised data already reveal. The paper shows predictability and DP are generally incomparable, though in the worst case (all but one record compromised, all binary queries sensitive) predictability implies mutual-information DP. The authors also build a general framework using the generalized method of moments (GMM) to analyze asymptotic predictability when compromised data come from stationary, ergodic, mixing processes, yielding a predictability-calibrated output perturbation scheme for empirical risk minimization that composes with DP.

Paper Overview

Field: ML Authors: Linda Lu, Karthik Sridharan Published: 2025-06-20 arXiv: 2506.16801

Abstract

Differential privacy (DP) ensures rigorous individual-level privacy guarantees against even the most knowledgeable attackers, but its worst-case nature can impose a costly privacy-accuracy tradeoff.

This paper introduces privacy via predictability, a fine-grained framework that explicitly incorporates:

  • the attacker's core knowledge,
  • a compromised portion of the dataset generated by a stochastic process, and
  • a specified family of queries.
  • Predictability measures privacy leakage as the incremental gain in an attacker's ability to predict sensitive information about unknown individuals after observing the algorithm's output, beyond what can already be inferred from the compromised data.

    The authors show that predictability and DP are generally incomparable: each can be small while the other is large. However, in the worst-case regime — all but one individual compromised, and all binary queries sensitive — predictability implies mutual-information DP. More broadly, predictability provides a finer-grained privacy measure tailored to specific sensitive information and specific attacker models.

    The authors also build a general framework using the generalized method of moments (GMM) to analyze asymptotic predictability when the compromised data come from stationary, ergodic, mixing processes. This yields a predictability-calibrated output perturbation scheme for empirical risk minimization (ERM), which complements DP and can be combined with it for fine-grained privacy control.

    Links

  • arXiv page: https://arxiv.org/abs/2506.16801

Tags

#differential-privacy#machine-learning#privacy#predictability#statistical-learning#arxiv-paper

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/177981554