Collaborative synthetic data improves fraud detection across companies
Collaborative Synthetic Data for Privacy-Preserving Financial Fraud Detection Across Organizational Silos
Machine LearningArtificial Intelligence
Summary
Financial fraud is hard to detect because fraud cases are rare, and data is spread across different organizations with privacy limits. The paper looks at a method called CollaFuse that creates fake data by combining information from multiple organizations without sharing real data. While this fake data isn't always perfectly realistic for each organization, it helps improve fraud detection more consistently than some traditional methods. This shows that sharing patterns across organizations is more important than making locally perfect data.
What this means in practice
- •For financial fraud analysts: Improve fraud detection models by using synthetic data generated collaboratively without sharing sensitive customer data.
- •For healthcare data teams: Generate synthetic patient data collaboratively across hospitals to improve rare disease detection while preserving privacy.
Authors
Simeon Allmendinger, Domenique Zipperling, Burhanettin Bahadir Kibar, Niklas K{ü}hl
Abstract
Organizations seek analytical value from AI, yet relevant data are often fragmented across organizations and constrained by privacy. This is acute in financial fraud detection, where rare fraud cases and imbalanced local datasets limit decision-relevant analytics. Federated learning enables collaboration without direct data sharing but does not resolve minority-class scarcity. Synthetic data generation can help, yet lightweight methods are interpolation-bound, while generative models require substantial data and computation. Existing collaborative generative approaches often rely on federated learning, imposing considerable organization-side training burdens. In this paper, we examine CollaFuse as a collaborative diffusion-based alternative for fraud detection and evaluate it across five fraud datasets. Compared with classical oversampling, local generative baselines, and centralized diffusion benchmarks, CollaFuse does not achieve the highest local fidelity but improves downstream fraud detection more consistently across most datasets. These findings suggest that synthetic data create analytical value less through local realism than through transferable cross-organizational structure.