Aug 2026· Data mining and knowledge discovery· Vol 40· 0 citations· 4 references
Computer Science
TL;DR
FedFairLAB is introduced, a FL method that enforces group, intersectional, and multiclass fairness simultaneously at both the local and global levels and a tunable performance budget allows practitioners to control how much predictive performance can be sacrificed to improve fairness.
Abstract
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.
Machine learning models are traditionally optimized for predictive accuracy, often overlooking critical aspects such as fairness and computational efficiency, which are essential for real-world deployment in socially sensitive and resource-constrained environments. This creates a significant research gap, as existing approaches typically address fairness or efficiency in isolation, lacking a unified framework that systematically balances multiple objectives. To address this limitation, this study proposes a multi-objective optimization framework that simultaneously integrates accuracy, fairness, and efficiency within the model development process using Pareto-based optimization techniques. The methodology involves training multiple machine learning models across benchmark datasets containing sensitive attributes, enabling the evaluation of trade-offs between objectives. The framework employs fairness metrics such as demographic parity and equal opportunity, alongside computational efficiency indicators including training time and resource utilization. Pareto front analysis is used to identify optimal model configurations that achieve balanced performance across competing criteria. The results demonstrate that the proposed approach achieves accuracy levels within 1–3% of the best-performing single-objective models, while reducing fairness disparities by up to 40% and computational cost by approximately 20–30%. Statistical analysis confirms that improvements in fairness and efficiency are significant (p < 0.01), with no statistically significant loss in accuracy. These findings highlight the effectiveness of multi-objective optimization in producing balanced and deployable machine learning systems. This study aims to advance a holistic optimization paradigm for responsible AI, enabling the development of models that are not only accurate but also fair and efficient, thereby aligning machine learning practices with ethical and operational requirements
Timur Dali Purwanto, M. Alshar'e, Anjali Bhardwaj et al.· Journal of Data Science· 0 citations
As machine learning (ML) algorithms are increasingly used in high-stakes applications, concerns have arisen that they may be biased against certain social groups. Although many approaches have been proposed to make ML models fair, they typically rely on the assumption that data distributions in training and deployment are identical. Unfortunately, this is commonly violated in practice and a model that is fair during training may lead to an unexpected outcome during its deployment. Although the problem of designing robust ML models under dataset shifts has been widely studied, most existing works focus only on the transfer of accuracy. In this paper, we study the transfer of both fairness and accuracy under domain generalization where the data at test time may be sampled from never-before-seen domains. We first develop theoretical bounds on the unfairness and expected loss at deployment, and then derive sufficient conditions under which fairness and accuracy can be perfectly transferred via invariant representation learning. Guided by this, we design a learning algorithm such that fair ML models learned with training data still have high fairness and accuracy when deployment environments change. Experiments on real-world data validate the proposed algorithm. Model implementation is available at https://github.com/pth1993/FATDM.
Machine learning has become increasingly prevalent in the world as a result of a combination of factors, notably computing breakthroughs and increased data availability. The three most common criteria of fairness in machine learning, despite their common sense and moral appeal, are often mutually exclusive. This frequently presents a challenge and the need to prioritize one criterion over the others. Specifically, the paper highlights the general overview of a machine learning model and presents an example of its application in the legal field that explores the existing biases in models today. We then delve into the three main criteria of machine learning fairness: independence (fairness based on outcomes being unrelated to different characteristics), separation (fairness based on equal error rates across groups), and sufficiency (fairness based on predictions being equally reliable for all groups). In the present experiment, we aimed to investigate the mutual exclusivity of these three machine-learning fairness criteria. We hypothesized that the fairness criteria being evaluated cannot all three be satisfied simultaneously, leading to a machine learning model that remains unfair. The results demonstrated that no threshold in the model simultaneously satisfied independence, separation, and sufficiency, highlighting the limitations of machine learning models that pose various issues across different sectors.
Keshav Pillutla, Kevin Fry· Journal of emerging investig...· 0 citations
Federated learning (FL) allows multiple institutions to train a shared model without exchanging raw data, which makes it attractive for privacy-sensitive domains such as finance. Deploying FL across jurisdictions, however, remains difficult: privacy regimes such as the General Data Protection Regulation (GDPR) in Europe and the California Consumer Privacy Act (CCPA) impose different, and sometimes conflicting, obligations on data handling. This paper proposes a regulatory-driven FL framework that treats compliance and fairness as first-class optimization objectives rather than as afterthoughts. The framework combines locally trained models with record-level differential privacy, secure aggregation of masked model updates, and a multi-objective loss that balances predictive accuracy, jurisdiction-specific compliance constraints, and group-fairness criteria, followed by a post-hoc bias audit that corrects residual disparities. We implement the complete pipeline and evaluate it on a synthetic cross-jurisdictional credit-scoring benchmark with three regions governed by GDPR-, CCPA-, and locally-styled privacy budgets, comparing against centralized learning and baseline federated averaging over five random seeds. The proposed framework keeps every region within its differential-privacy budget (ε = 1.58, 2.41, and 3.32 against caps of 2.0, 3.0, and 4.0 at δ = 10⁻⁵), raises the disparate impact ratio from 0.88 to 0.99, reduces the demographic-parity difference from 0.068 to 0.007, and attains a perfect score on an operational compliance rubric, while giving up only 0.9 percentage points of accuracy (88.2% versus 89.1%) at roughly 2.6 times the training cost. The results also surface a known tension: enforcing demographic parity increases the equalized-odds gap, quantifying the price of fairness under differential privacy in federated financial systems.
J. Nalavade· Journal of Intelligent Decis...· 0 citations
Many real-world systems rely on predictive models to inform decisions, and fairness concerns arise in both the prediction and decision stages. We introduce end-to-end fairness optimization (E2EFO) as a unifying framework that integrates fairness across the prediction-to-decision pipeline. We focus on resource allocation with group-based fairness: the prediction task estimates allocation impacts while limiting accuracy disparity across groups, and the decision task distributes those impacts equitably by optimizing a group-based alpha-fairness measure. Within this framework, we propose fair decision-focused learning (FDFL), a training paradigm that jointly accounts for prediction accuracy, prediction fairness, and decision regret -- the loss in decision fairness due to imperfect predictions. FDFL trains the predictor by gradient descent, combining the objective gradients through multi-task learning techniques. The core computational challenge is the decision Jacobian with respect to the predictor parameters: we derive exact closed-form formulas for a tractable class of fair allocation and apply a differentiable optimization layer in the general case. We further establish a finite-sample generalization bound for the scalarized FDFL objective. Numerical experiments on a healthcare-based single resource allocation and a synthetic multiple resource allocation illustrate the value of jointly accounting for prediction fairness and decision fairness in prediction-informed decision-making.
Accurate prediction of ICU mortality is important for clinical decision-making and resource allocation. In the Federated Learning (FL) approach, privacy is maintained using decentralized hospital data to train machine learning models. Biases in hospital data may increase health inequalities. To address the issue, a new fairness-aware FL approach is developed. The two main metrics, Equalized Odds and Demographic Parity, are integrated in the FL optimization process. The proposed FL model uses adaptive client weighting, thereby minimizing bias propagation within the model. The approach is tested using the MIMIC-IV database. The experimental results demonstrate that the proposed model reduces subgroup differences without compromising predictive accuracy. Ablation studies prove that the combination of fairness regularization and client weighting gives better results. Statistical analysis validates that the achieved fairness improvement is clinically meaningful and robust. The findings conclude that fairness-aware federated learning for predicting critical care outcomes is a strong and viable framework balancing privacy, accuracy, and fairness. This proves to be an important step in the development of ethical AI in healthcare.
A weeklong summer workshop brought higher education faculty to campus to explore how AI and machine learning materials can be adapted for their classrooms.
