A Dual-Path Generative Framework for Zero-Day Fraud Detection in Banking Systems

TL;DR

This paper introduces a dual-path generative framework combining real-time anomaly detection with offline adversarial training to improve zero-day fraud detection in banking systems, balancing low latency and explainability.

Contribution

It proposes a novel architecture that decouples detection and training, using VAE and WGAN-GP, with a Gumbel-Softmax estimator and trigger-based SHAP explanations for effective zero-day fraud detection.

Findings

Achieves <50ms inference latency for real-time detection

Effectively synthesizes fraudulent scenarios for stress-testing

Provides explainability selectively for high-uncertainty transactions

Abstract

High-frequency banking environments face a critical trade-off between low-latency fraud detection and the regulatory explainability demanded by GDPR. Traditional rule-based and discriminative models struggle with "zero-day" attacks due to extreme class imbalance and the lack of historical precedents. This paper proposes a Dual-Path Generative Framework that decouples real-time anomaly detection from offline adversarial training. The architecture employs a Variational Autoencoder (VAE) to establish a legitimate transaction manifold based on reconstruction error, ensuring <50ms inference latency. In parallel, an asynchronous Wasserstein GAN with Gradient Penalty (WGAN-GP) synthesizes high-entropy fraudulent scenarios to stress-test the detection boundaries. Crucially, to address the non-differentiability of discrete banking data (e.g., Merchant Category Codes), we integrate a Gumbel-Softmax estimator. Furthermore, we introduce a trigger-based explainability mechanism where SHAP (Shapley Additive Explanations) is activated only for high-uncertainty transactions, reconciling the computational cost of XAI with real-time throughput requirements.