The paper introduces a shift‑aware score transport method for privacy auditing that does not require retraining the target model. It estimates the in‑distribution non‑member score distribution from out‑of‑distribution data, transports this relationship to the target covariate distribution, and applies a conservative correction to account for potential extrapolation errors. Experiments on multiple datasets with distribution shifts show that the method provides more stable privacy estimates while maintaining tighter lower bounds than existing baselines.
arXiv:2606. 16952v2 Announce Type: replace-cross Abstract: The rapid adoption of generative AI and Large Language Models (LLMs) has spurred interest in synthetic data as a privacy-preserving alternative to sensitive real-world datasets.
arXiv:2606. 16952v1 Announce Type: cross Abstract: The rapid adoption of generative AI and Large Language Models (LLMs) has spurred interest in synthetic data as a privacy-preserving alternative to sensitive real-world datasets.
The paper addresses a flaw in differential privacy audits where selection bias in Membership Inference Attacks leads to overestimated lower bounds on privacy leakage. By applying statistical correction methods, the authors demonstrate that uncorrected audits can violate theoretical privacy guarantees up to four times more often than claimed, while corrected audits maintain valid bounds at the nominal confidence level.
arXiv:2602. 02819v4 Announce Type: replace Abstract: Membership Inference Attacks (MIAs) aim to distinguish training points (members) from unseen data (non-members), and are widely used to quantify memorization and assess privacy risks.
arXiv:2605. 27292v2 Announce Type: replace Abstract: Privacy auditing aims to empirically assess privacy leakage in machine learning models using membership inference attacks (MIAs), and to derive lower bounds on differential privacy (DP) parameters.
arXiv:2608.28934v1 Announce Type: new Abstract: Differential privacy (DP) has traditionally been used to provide theoretical upper bounds on an algorithm's stability to changing its training data. In...
arXiv:2606. 10481v1 Announce Type: cross Abstract: Parameter-efficient fine-tuning of large language models (LLMs) can exhibit problematic memorization of individual training examples.
The paper introduces Pairwise Likelihood MIA (PL‑MIA), a unified membership inference attack that combines a Gaussian likelihood‑ratio statistic with population calibration and the Cauchy combination test. PL‑MIA generates p‑values from pairwise comparisons between a query point and reference points, then aggregates these continuous signals using the Cauchy test to preserve evidence strength. Experiments show that PL‑MIA surpasses strong baselines, boosting true positive rates by over 25% in low‑false‑positive settings, thereby validating the theoretical advantages of the proposed statistical framework.
arXiv:2606. 24408v1 Announce Type: new Abstract: Assessing the privacy of large language models (LLMs) presents significant challenges.
The paper introduces IBIS, a framework for designing canaries that are efficient for one‑run privacy auditing. IBIS optimizes each canary’s membership inference attack score while minimizing interference among canaries by encouraging orthogonality in embedding space. The method uses influence‑based greedy initialization and a bilevel optimization algorithm that updates a single model incrementally, achieving better performance‑cost trade‑offs and reducing computational cost by up to two orders of magnitude compared to existing approaches.
The paper introduces a subgroup-targeted membership inference game to audit differentially private synthetic text releases, revealing that existing average-case attacks miss significant leakage to vulnerable subgroups. An extensive audit across 32 proxies, four datasets, three generation methods, and five privacy budgets shows that DP reduces overall leakage but leaves concentrated, uneven residual risk, especially for high-risk records. The study demonstrates that which records leak is determined by the release mechanism rather than the records themselves, challenging record-level risk assessment.
arXiv:2606. 17464v1 Announce Type: new Abstract: Membership inference attacks (MIAs) are a canonical way to assess a machine learning model's privacy properties.