Locally Differentially Private Embedding Models in Distributed Fraud Prevention Systems
This addresses the challenge of data sharing in fraud prevention for financial institutions, though it appears incremental by applying known privacy techniques to a specific domain.
The paper tackles the problem of enabling collaborative fraud detection across financial institutions without sharing sensitive data by introducing a locally differentially private embedding model, and demonstrates its robustness to inference attacks and utility-privacy trade-offs on real payment network datasets.
Global financial crime activity is driving demand for machine learning solutions in fraud prevention. However, prevention systems are commonly serviced to financial institutions in isolation, and few provisions exist for data sharing due to fears of unintentional leaks and adversarial attacks. Collaborative learning advances in finance are rare, and it is hard to find real-world insights derived from privacy-preserving data processing systems. In this paper, we present a collaborative deep learning framework for fraud prevention, designed from a privacy standpoint, and awarded at the recent PETs Prize Challenges. We leverage latent embedded representations of varied-length transaction sequences, along with local differential privacy, in order to construct a data release mechanism which can securely inform externally hosted fraud and anomaly detection models. We assess our contribution on two distributed data sets donated by large payment networks, and demonstrate robustness to popular inference-time attacks, along with utility-privacy trade-offs analogous to published work in alternative application domains.