Unsupervised quantum machine learning for fraud detection

TL;DR

This paper explores quantum kernel methods for credit card fraud detection, demonstrating that quantum approaches can outperform classical benchmarks as the number of features increases, with promising results up to 20 qubits.

Contribution

The paper introduces quantum kernel-based anomaly detection protocols for fraud detection and shows their advantage over classical methods with increasing system size.

Findings

Quantum kernels with re-uploading outperform classical methods at larger feature sizes.

At 20 qubits, quantum methods achieve a 15% higher average precision.

Quantum protocols show potential for near- and mid-term hardware implementation.

Abstract

We develop quantum protocols for anomaly detection and apply them to the task of credit card fraud detection (FD). First, we establish classical benchmarks based on supervised and unsupervised machine learning methods, where average precision is chosen as a robust metric for detecting anomalous data. We focus on kernel-based approaches for ease of direct comparison, basing our unsupervised modelling on one-class support vector machines (OC-SVM). Next, we employ quantum kernels of different type for performing anomaly detection, and observe that quantum FD can challenge equivalent classical protocols at increasing number of features (equal to the number of qubits for data embedding). Performing simulations with registers up to 20 qubits, we find that quantum kernels with re-uploading demonstrate better average precision, with the advantage increasing with system size. Specifically, at 20 qubits we reach the quantum-classical separation of average precision being equal to 15%. We discuss the prospects of fraud detection with near- and mid-term quantum hardware, and describe possible future improvements.