Mitigating Explanation Leakage in Financial Fraud Detection Systems

Muhammad Waleed Gul; Elaheh Homayounvala

cs.LGAug 23, 2026

Mitigating Explanation Leakage in Financial Fraud Detection Systems

Authors: Muhammad Waleed Gul, Elaheh Homayounvala

Organizations: MSc Artificial Intelligence London Metropolitan University London, United Kingdom · School of Computing and Digital Media London Metropolitan University London, United Kingdom

Abstract

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).

Figures & tables

Architecture	AUPRC	MIA Acc.	Top-10	Spearman
Centralized Baseline	$0.9012\pm 0.0031$	61.67%	100.0%	1.000
Standard Federated	0.8989 *	51.58%	100.0%	0.980
DP-FedSHAP ( $\epsilon=1.2$ )	0.8989 *	49.75%	80.0%	0.793
Weight-Level DP ( $\epsilon=1.2$ )	0.5040 *	49.92%	70.0%	0.7623

TABLE I: Privacy–Utility–Fidelity Trade-off Across Architectures

Explore similar work

Enhanced Detective: A Dual-Stage Proactive Defense Against Internal Poisoning Attacks in Financial Fraud Detection Using Federated Learning
Oct 13, 2026Maria-Ioana-Andrea Ivan, Orhan Ermis, Ruxandra F. OlimidFederated Learning Aggregation Federated Learning
Department of Computer Science, University of Bucharest · Luxembourg Institute of Science and Technology (LIST), Luxembourg · Research Institute of the University of Bucharest (ICUB), Bucharest, Romania
Privacy-Preserving Federated Fraud Detection in Payment Transactions with NVIDIA FLARE
Mar 13, 2026Holger R. Roth, Sarthak Tickoo, Mayank Kumar +18Financial Fraud Detection Anomaly Detection
NVIDIA, Santa Clara, USA · J.P. Morgan, New York, USA · DeepTempo, San Francisco, USA +2
DPxFin: Adaptive Differential Privacy for Anti-Money Laundering Detection via Reputation-Weighted Federated Learning
Mar 14, 2026Renuga Kanagavelu, Manjil Nepal, Ning Peiyan +6Differentially Private Federated Learning Federated Learning
A*STAR IHPC Singapore · SRM University-AP Andhra Pradesh, India · EVYD Singapore

Oct 13, 2026J/K move · Enter open · S save

Enhanced Detective: A Dual-Stage Proactive Defense Against Internal Poisoning Attacks in Financial Fraud Detection Using Federated Learning

Maria-Ioana-Andrea Ivan, Orhan Ermis, Ruxandra F. Olimid

Department of Computer Science, University of Bucharest · Luxembourg Institute of Science and Technology (LIST), Luxembourg · Research Institute of the University of Bucharest (ICUB), Bucharest, Romania

Federated Learning Aggregation Federated Learning Byzantine-Robust Federated Learning Financial Fraud Detection

Open paper AskPDF arXiv

With recent developments in Artificial Intelligence (AI), the adoption of distributed learning paradigms such as Federated Learning (FL) has emerged. In simple terms, FL enables multiple entities to collaboratively train AI models without sharing local data. The FL paradigm is particularly valuable in privacy-sensitive settings, such as financial services, because it enables institutions to collaboratively improve fraud-detection models while keeping sensitive transaction data local. However, FL schemes are prone to security vulnerabilities, where malicious clients can poison local data or model updates and degrade the global model, and hence the performance of the resulting model might be affected. Therefore, in this work, we propose Enhanced Detective, a dual-stage robust aggregation strategy that combines Median Absolute Deviation (MAD), distance-based update filtering, and FedMedian aggregation. Experiments on an imbalanced credit card dataset evaluate targeted label-flipping and weight-scaling attacks over 50 communication rounds. With 40% malicious clients, Enhanced Detective achieves an AUPRC of 0.796 and a Recall of 0.847.

Abstract

Explore similar work

Enhanced Detective: A Dual-Stage Proactive Defense Against Internal Poisoning Attacks in Financial Fraud Detection Using Federated Learning

Privacy-Preserving Federated Fraud Detection in Payment Transactions with NVIDIA FLARE

DPxFin: Adaptive Differential Privacy for Anti-Money Laundering Detection via Reputation-Weighted Federated Learning