joseph gabriel ragoncin
news editor
Archer has completed the latest phase of its quantum machine learning fraud detection project, which includes testing on both quantum simulators and commercial quantum computers.
Its quantum neural network model was benchmarked on a public financial fraud dataset containing over 280,000 transaction records. The model was designed to work within current quantum computing limitations using dimensionality reduction and data balancing techniques before testing.
In the simulator environment, the initial model matched the most powerful classical model in the Archer benchmark. It correctly identified 118 fraudulent transactions, missed 30, and generated 1 false positive.
False positives are a key issue in fraud detection because they can lead to unnecessary reviews and customer confusion. Archer highlighted the simulator’s low false positive rate as one of the most notable results from this phase of the project.
The model was also tested under simulated quantum noise. The system was stable at low noise levels and showed only slight degradation at moderate levels, but performance degraded more rapidly at higher noise levels.
This gives Archer a clearer understanding of the hardware requirements needed to implement quantum machine learning in fraud inspection. It also reduces technical uncertainties regarding how such models will perform outside of ideal test conditions.
Another validation was performed on IQM Garnet, a 20-qubit superconducting quantum computer accessed via AWS Braket. In that hardware test, the model detected 18 out of 19 fraudulent transactions in the test set.
Archer said running real hardware resulted in a higher false positive rate than simulator testing. Still, he said the results are evidence that the model can work not only in simulations but also on current commercially available quantum hardware.
benchmark work
Archer used a step-by-step experimental framework to determine model configuration. This process includes studying qubit selection, optimizing feature maps, comparing with classical machine learning methods, and analyzing the effects of quantum noise.
According to Archer, this research established a reproducible benchmarking framework and identified a quantum architecture that performs well on selected fraudulent datasets. He also said that the project has not yet shown clear advantages over state-of-the-art classical artificial intelligence methods.
Archer’s progress comes as banks and quantum computing groups continue to consider fraud detection as one of the more practical commercial uses of quantum machine learning. The report pointed to Quantinuum’s work with HSBC and separate collaborations with Intesa Sanpaolo and IBM as examples of similar efforts elsewhere in the market.
These programs reflect broader interest by financial institutions in tools that can analyze large volumes of transactions while limiting missed fraud and unnecessary alerts. For quantum computing companies, fraud detection provides predefined test cases with measurable results for existing machine learning systems.
research stage
The latest milestone follows Archer’s early completion of dataset preparation before the project moves into simulation and benchmarking. This phase was performed using a prepared research dataset and a selected set of comparison models rather than production banking conditions.
More testing is needed to evaluate the commercial path. Archer said this includes large datasets, additional classical benchmarks, iterative trials, and further hardware validation.
This research forms part of Archer’s broader activities in quantum computing, sensing and medical diagnostics. The company operates in the semiconductor field and has developed chips related to quantum applications.
Dr. Simon Ruffell, CEO of Archer, said the results were a useful step in testing whether quantum machine learning can be applied to fraud detection under current technological limitations.
“These results demonstrate that the QML approach can provide strong fraud detection performance while operating within the constraints of current quantum computing systems. The simulator results are robust, especially with very low false positive rates, and successful execution on real quantum hardware is an important validation step. CSIRO “The research collaboration agreement with QML forms part of Archer’s strategy to explore the practical application of quantum computing technology and support future commercialization opportunities in data-intensive industries. Fraud detection is a relevant use case for QML and payment providers need to quickly analyze large amounts of transaction data while reducing both missed fraud and false alarms.”
