Jul 2026· International Conference on Information and Communicatiaon Technology· pp. 1-6· 0 citations· 13 references
Abstract
Federated learning allows financial institutions to collaboratively identify fraud without distributing raw transaction data, while differential privacy safeguards individual records from inference attacks. Utilizing a lightweight four-layer neural network that was trained on a 10,000-sample subset of the PaySim mobile money dataset (originally 6.3 million transactions), this paper provides a systematic empirical evaluation of a differentially private federated fraud detection framework. The framework is evaluated in the context of varying data distributions (IID and multiple non-IID schemes), clients' numbers, participation rates, privacy budgets, and data quantity skews. The findings indicate that IID and moderately heterogeneous settings produce high accuracy and F1-scores. Conversely, performance is substantially undermined by severe non-IID partitions, numerous small clients, and extreme data imbalances, particularly when differential privacy noise is implemented. The non-private model's detection capability is largely preserved by intermediate privacy budgets (approximately ε≈1). Notably, membership inference attack success rates remain at or near random guessing (approximately 50% accuracy) in the absence of differential privacy, indicating that the baseline provides a restricted attack surface. In this evaluation setting, the privacy benefit is further confirmed by the fact that differential privacy at ε≈1 further suppresses the attack advantage toward zero. While these results provide practical configuration advice for federated, privacy-preserving fraud detection systems, they should be interpreted within the context of the simplified experimental setup that was implemented.
Aim/Purpose
This study investigates whether federated machine learning (FML) can match or exceed centralized machine learning (CML) performance for financial fraud detection while maintaining complete data privacy.
Background Financial fraud detection systems traditionally rely on CML, which aggregates sensitive transaction data from multiple institutions, raising privacy concerns and regulatory compliance challenges under the GDPR and CCPA.
Methodology
Using the Kaggle Credit Card Fraud dataset (284,807 transactions), this study compared identical neural network architectures in centralized and federated settings. The FML approach utilized the FedAvg algorithm across 10 clients with non-IID data distribution over 20 rounds.
Contribution
This research challenges the assumption that centralized learning out-performs federated approaches, providing the first comprehensive empirical comparison showing FML superiority on real-world fraud data.
Findings
FML outperformed CML in F1-score (0.7957 vs. 0.7411) and precision (0.8409 vs. 0.6587). While FML significantly reduces false positives and operational costs, CML remains competitive when fraud losses outweigh the costs of false positives.
Recommendations for Practitioners
Financial institutions should adopt FML for fraud detection when multi-institution collaboration or data-localization compliance is required. Its higher precision makes it well-suited for production environments focused on reducing false positives.
Recommendation for Researchers
Future research should examine advanced aggregation, differential privacy, and scalability. It should also analyze regularization effects to determine why FML achieves superior precision.
Impact on Society
This research shows that privacy preservation can enhance fraud detection, enabling multi-institution collaboration while protecting consumer data and reducing friction by reducing false positives.
Future Research
Future work should explore FML in real-time institutional settings, across diverse fraud contexts, and within integrated cross-domain frameworks such as credit risk assessment.
Samuel Sambasivam· Issues in Informing Science...· 0 citations
Federated learning is a decentralised machine-learning approach in which several clients jointly build a shared model without moving their raw data to one location. Rising concerns around privacy, tightening regulation, and restrictions on how data may be owned or shared have made this approach increasingly attractive in practice. Although federated learning lowers privacy exposure relative to centralised training, deploying it in practice is complicated by clients whose data are unevenly distributed and non-identically distributed, by clients that participate inconsistently, and by training that can converge unpredictably. To obtain global models that train reliably and consistently even when client data are heterogeneous, this work puts forward a federated learning system built around privacy preservation. The design follows a client–server pattern in which a coordinating server aggregates updates from local models using weights that account for imbalance among participants. The behaviour of the resulting system is examined methodically across several data-distribution regimes — IID, mildly non-IID, and severely non-IID. The experiments show that the framework converges reliably and delivers predictive accuracy that holds up well, especially in the more difficult non-IID cases. Compared with conventional federated learning baselines, the approach shows greater robustness and steadier performance across successive training rounds. Because it is simple to implement, repeatable, and built with real deployment in mind, the architecture suits privacy-sensitive, decentralised use cases such as distributed intelligent systems, industrial monitoring, and healthcare analytics.
Shyam Patel, S. Khan· 2026 International Conferenc...· 0 citations
SecureFedShield is proposed, a privacy-preserving federated learning framework designed for secure financial fraud detection in adversarial environments that integrates adaptive privacy protection, trust-aware client evaluation, adversarial update detection, and robust model aggregation into a unified architecture.
Kriti Mishra· International Journal of Cre...· 0 citations
This study implemented a comprehensive experimental framework for analysing FL performance using standard FL aggregation protocols FedAvg, FedProx, and SCAFFOLD in conjunction with Differential Privacy mechanisms; specifically, the Gaussian noise mechanism with Rényi Differential Privacy (RDP) accountants.
Financial fraud detection relies heavily on centralized machine learning models. This creates serious data privacy risks. Federated Learning (FL) decentralizes data processing, but financial regulations still require models to be transparent. This means using Explainable AI (XAI) tools such as TreeSHAP. Recent cybersecurity research shows a problem with this approach. Sharing high-fidelity SHAP explanations exposes the federated network to Membership Inference Attacks (MIAs). This dissertation proposes and evaluates DP-FedSHAP. It is a new architecture that applies client-level differential privacy only to post-hoc TreeSHAP vectors. It is compared against a Weight-Level DP baseline, which perturbs the trained model directly instead. Using the highly imbalanced IEEE-CIS Fraud Detection dataset, this study measures the trade-off between explanation fidelity, privacy preservation, and the model's Area Under the Precision-Recall Curve (AUPRC).
The 340B Drug Pricing Program, critical for supporting safety-net healthcare providers, faces significant challenges from fraud, waste, and abuse, particularly through drug diversion and duplicate discounts. Current program integrity relies heavily on static, rule-based audits, which are increasingly ineffective against novel and sophisticated, collusive fraud schemes. This paper proposes a novel, multi-layered artificial intelligence (AI) framework designed to move beyond this reactive paradigm. The framework combines: (1) unsupervised transactional anomaly detection, utilizing Autoencoders and Isolation Forests to identify entity-level outliers; and (2) relational fraud detection, using an inductive Graph Neural Network (GNN), GraphSAGE, to model the complex relationships between covered entities, contract pharmacies, and providers to uncover collusive networks. Crucially, this hybrid model is designed within a privacy-preserving architecture. We propose the use of Federated Learning (FL) and Differential Privacy (DP) to train the anomaly detection modules collaboratively, allowing participating entities to build robust, global models without centralizing or exposing sensitive HIPAA-protected patient data. The framework integrates Explainable AI (XAI) methodologies (SHAP and GNNExplainer) for regulatory auditability and a concept drift detection component with a Human-in-the-Loop (HITL) active learning loop to ensure long-term adaptability to evolving fraud tactics. This approach provides a blueprint for a proactive, adaptive, and privacy-compliant system to safeguard 340B program integrity.
Pinaki Bose· International journal of com...· 0 citations