Jul 2026· Proceedings on Privacy Enhancing Technologies· Vol 2026, pp. 53-82· 0 citations· 58 references
Computer Science
TL;DR
It is demonstrated that while DP alone can degrade both utility and fairness, applying fairness interventions can partially restore equitable outcomes, and post-processing methods tend to provide more stable fairness–utility trade-offs across privacy budgets and synthesizers.
Abstract
Machine learning models are increasingly deployed in high-stakes domains, raising concerns about both privacy and fairness. Differential Privacy (DP) has become a gold standard for privacy-preserving data analysis, while fairness-aware mechanisms aim to mitigate discrimination against underrepresented groups. However, these objectives can conflict: DP often amplifies disparities across demographic groups, and little is known about whether established fairness interventions remain effective under DP constraints. In this work, we present, to our knowledge, the first systematic evaluation of fairness interventions on differentially private synthetic tabular data. Our benchmark centers on the Adaptive Iterative Mechanism (AIM), identified as the state-of-the-art marginal-based DP synthesizer in recent KDD & VLDB 2025 tutorials by Cormode et al. We thus evaluate fairness interventions across four datasets, multiple group fairness metrics, and three categories of mitigation strategies (pre-processing, in-processing, and post-processing) under a wide range of privacy budgets. We compare four pipeline configurations: (Baseline) training on original data; (DP-only) training on DP synthetic data; (Fair-only) applying fairness mechanisms on original data; and (DP+Fair) combining fairness mechanisms with DP synthetic data. Our results demonstrate that while DP alone can degrade both utility and fairness, applying fairness interventions can partially restore equitable outcomes. Among them, post-processing methods tend to provide more stable fairness–utility trade-offs across privacy budgets and synthesizers, achieving strong fairness improvements while preserving competitive utility relative to other intervention stages. We release all code, data, and experimental artifacts in an open-source repository (https://github.com/vinicius-verona/dp-fair-intervention-benchmark) to ensure full reproducibility and to support future research on the privacy-fairness-utility trade-off.
This work argues that privacy cost, the information leakage borne by each group, is itself a form of harm, and adopts a compensatory-fairness framework in which a group that involuntarily bears greater privacy exposure is owed proportionally greater benefit from the system.
Machine learning (ML) models deployed in sensitive domains such as healthcare, law enforcement, and finance must satisfy not only utility requirements but also fairness and privacy guarantees. While prior work has largely examined how privacy-preserving techniques affect fairness, the inverse question-how fairness-enhancing algorithms influence privacy leakage-remains underexplored. We present the first comprehensive study of how fairness interventions affect membership inference privacy risks at the subpopulation level. By adapting the Likelihood Ratio Attack (LiRA) for subgroup auditing, we uncover privacy disparities that aggregate evaluations obscure. We further analyze how Differential Privacy (DP) interacts with fairness-enhancing methods across different categories, showing that DP’s privacy benefits and utility costs are unevenly distributed across subpopulations. Our results demonstrate that fairness interventions do not uniformly increase privacy risk; their impact depends on model architecture, subgroup size, and mitigation strategy. These findings reveal that fairness, privacy, and utility must be jointly evaluated at the subpopulation level, and we introduce the first unified empirical framework to support such auditing in practice.
U. Suleymanov, Ilhama Novruzova, Khalid Mammadov et al.· European Symposium on Securi...· 0 citations
Machine learning increasingly drives decisions in domains such as finance and healthcare, where ethical considerations, such as fairness, are central. In such contexts, ensuring fairness is essential, especially when decisions impact individuals and social groups. Federated learning (FL) provides a decentralized training paradigm, yet client heterogeneity and demographic imbalance can amplify disparities across subpopulations. Existing fairness-aware FL methods remain limited, often focusing on group fairness in binary classification and lacking explicit control over the trade-off between fairness and predictive performance. We introduce FedFairLAB, a FL method that enforces group, intersectional, and multiclass fairness simultaneously at both the local and global levels. A tunable performance budget allows practitioners to control how much predictive performance can be sacrificed to improve fairness. Experiments on six real-world datasets show that FedFairLAB substantially improves fairness while keeping models accurate and usable in realistic FL settings.
Michele Fontana, Francesca Naretto, A. Monreale· Data mining and knowledge di...· 0 citations
Deep learning models often produce performance disparities across demographic groups, due to the training data imbalance with respect to sensitive attributes such as gender or age. To address this problem, existing work has explored fair representation learning, data re-sampling, and adversarial training, which can be broadly categorized into two main approaches. Single-stage methods typically learn a shared representation for fairness, but often struggle to handle heterogeneous subgroup distributions. Two-stage methods learn representations separately from the final prediction task, which can lead to misalignment between fairness objectives and downstream predictions. We identify routing-induced bias, a failure mode in which subgroup imbalance drives the gating network to route subgroups onto a few experts, and propose an end-to-end Mixture-of-Experts (MoE) framework that corrects it. Specifically, we apply subgroup reweighting to correct data imbalance, and introduce gate entropy regularization to prevent routing from collapsing onto subgroup attributes, keeping expert utilization both balanced and interpretable. Beyond improving fairness, the routing distribution offers an interpretable view of how subgroups are allocated across experts. Experimental results demonstrate that the proposed approach improves fairness while maintaining competitive predictive performance.
Algorithmic fairness methods are increasingly used to identify and mitigate bias in machine learning models, yet most approaches are evaluated in isolation and along single demographic axes. This limits practical guidance for selecting fairness strategies, where disparities may arise across intersectional subgroups and across multiple stages of the modeling lifecycle. This work presents FairSelect, a toolkit for systematically evaluating fairness mitigation strategies applied individually and in combination across preprocessing, inprocessing, and postprocessing stages. FairSelect supports multiple model architectures, intersectional subgroup evaluation, and comparison of fairness utility tradeoffs across baseline, single method, and multi level configurations. The framework was validated using synthetic clinical datasets designed to represent specific bias mechanisms and a real-world replication of two-year stroke risk prediction among patients with atrial fibrillation. Synthetic experiments showed that targeted fairness methods generally reduced intended subgroup disparities, while combined strategies produced larger average fairness improvements with modest utility tradeoffs. In the clinical prediction task, mitigation effects were highly variable, with some combinations improving both fairness and predictive performance while others were ineffective or counterproductive. These findings demonstrate that fairness interventions interact in nonadditive and context dependent ways. FairSelect provides a practical framework for systematically identifying fairness strategies that improve subgroup equity while preserving model performance in clinical machine learning.
Nick Souligne, Isabella Mixton-Garcia, V. Subbian· 0 citations
Federated learning (FL) is a promising approach for training distributed machine learning models while preserving clients’ data privacy. However, in real-world FL systems, data are often not independent and identically distributed (non-IID). This heterogeneity can slow convergence, degrade model performance, and increase client drift. To address these challenges, numerous methods have been proposed to mitigate non-IID data effects by optimizing client selection, local training, and model aggregation strategies. Despite their effectiveness in improving performance and efficiency, these methods rarely consider fairness across clients. Improving global accuracy does not guarantee balanced participation, influence, or outcomes, which may lead to biased model behavior across clients. In this survey, we review existing non-IID mitigation methods in FL from a fairness perspective and provide a systematic analysis of their implicit impact on client participation and influence. Unlike prior surveys that treat fairness as a separate research direction, this work analyzes how these methods designed for non-IID mitigation implicitly shape fairness outcomes across clients. Our taxonomy classifies existing methods into three categories—fairness-aware, semi-fairness-aware, and fairness-unaware—based on their design strategies for client selection and model aggregation. Using this taxonomy, we analyze the advantages, trade-offs, and limitations of each category and highlight that mitigating non-IID data does not guarantee fairness across clients. Finally, we identify open challenges and outline future directions, including system-level FL design that jointly considers non-IID mitigation and fairness and the development of standardized fairness evaluation metrics. Overall, this survey aims to provide a structured perspective on the relationship between non-IID mitigation and fairness and support the development of more balanced and scalable FL systems under non-IID conditions.
Mohannad Alsofyani, Isra Al-Turaiki, H. Mathkour· Electronics· 0 citations