Quantum Computing and AI Combine to Enhance Payment Fraud Detection

AdvanThink, a European specialist in real-time payment fraud detection, and Quandela, a quantum computing company, have partnered to integrate quantum machine learning algorithms into existing fraud prevention models. The collaboration, announced on 17 June 2025 at Paris-Saclay, aims to improve the speed and accuracy of fraud detection systems currently reliant on artificial intelligence. Initial work has focused on integrating Quandelas existing quantum machine learning model – previously applied to credit risk assessment – into AdvanThink’s operational pipeline, with the intention of benchmarking performance against current industry standards and developing deployable solutions for financial institutions.

Partnership for Quantum Fraud Detection

AdvanThink and Quandela are collaborating to integrate quantum machine learning algorithms into existing payment fraud detection systems, aiming to enhance performance in speed, accuracy, and resilience against increasingly complex attacks. This partnership centres on developing a proof of concept to demonstrate the value of this integration within AdvanThink’s established industrial pipelines, facilitating a direct assessment of quantum capabilities. The demonstrator provides a platform for experts to evaluate the potential benefits of quantum computing in addressing the challenges of payment fraud. This successful integration demonstrates the feasibility of deploying quantum-enhanced fraud detection systems within a production environment.

Quandelas pre-existing quantum machine learning model, initially developed for credit risk assessment, is being adapted for quantum fraud detection, allowing for direct benchmarking against current market-leading products. Integrating this model into an industrial workflow facilitates a comparative analysis of performance metrics and identifies areas for improvement, providing valuable insights into quantum advantages.

Integrating Quantum Machine Learning

The integration of quantum machine learning algorithms exploits quantum phenomena to accelerate and refine pattern recognition within the large datasets characteristic of transaction monitoring, offering a potential leap in analytical power. Quandelas approach utilises a pre-trained quantum model, validated for credit risk, and adapts its computational structure to the specific demands of identifying fraudulent transactions, optimising performance on fraud-related feature sets. This adaptation involves re-parameterising the quantum circuit to ensure efficient and accurate fraud detection, leveraging the strengths of quantum computation.

Benchmarking against established fraud detection systems focuses on key performance indicators including detection rates for novel fraud schemes, false positive rates, and computational efficiency, measured in terms of processing time and energy consumption. This comparative analysis quantifies the extent to which quantum machine learning can surpass the capabilities of classical algorithms in a real-world operational context, providing concrete evidence of its potential. AdvanThink’s pipeline integration demonstrates a pathway for hybrid quantum-classical computation, where quantum processors handle computationally intensive tasks like feature extraction and model training.

A key aspect of the implementation addresses the challenges of interfacing quantum processors with existing classical computing infrastructure, enabling seamless integration into existing systems. Classical systems manage data pre-processing, post-processing, and overall system control, creating a practical and efficient workflow. The successful integration also necessitates careful consideration of data encoding strategies, converting classical transaction data into a quantum-compatible format while preserving relevant information and minimising quantum resource requirements.

Quandelas model employs techniques to optimise data transfer rates between classical and quantum processors, implementing efficient parallel processing strategies and developing robust error mitigation techniques to ensure data integrity. While the initial demonstrator showcases the feasibility of integrating quantum machine learning, expanding the system to handle the transaction volumes of a major financial institution requires careful architectural planning. Scalability remains a key consideration, demanding innovative solutions to manage increasing data loads and maintain system performance.

To mitigate limitations imposed by coherence times and qubit complexity, research focuses on developing quantum algorithms that achieve substantial performance gains with relatively modest quantum resources. This includes exploring variational quantum algorithms and quantum approximate optimisation algorithms, designed to operate efficiently on near-term quantum devices. Beyond performance gains, the integration of quantum computing offers potential benefits in terms of energy efficiency, reducing the computational intensity of complex machine learning models.

Classical machine learning models, particularly deep neural networks, can be computationally intensive and require substantial energy consumption, creating a significant environmental impact. Quantum algorithms, by leveraging the principles of quantum mechanics, may offer the possibility of achieving comparable or superior performance with significantly reduced energy requirements, promoting sustainable computing solutions. The partnership’s long-term vision includes exploring the application of quantum-enhanced fraud detection to other areas of financial crime, such as anti-money laundering and terrorist financing.

The underlying principles of quantum machine learning – accelerated pattern recognition and improved data analysis – are broadly applicable to a wide range of security challenges, extending its potential impact beyond fraud detection. By establishing a foundation for quantum-enhanced fraud detection, AdvanThink and Quandela aim to pave the way for a more secure and resilient financial ecosystem, safeguarding against evolving threats and ensuring financial stability.