Itaú and QC Ware use quantum computing principles to strengthen customer retention
Quantum algorithm for finance identifies customers at risk in banking, and defines future performance and accuracy advantages of performance and accuracy; model involved the use of two years of anonymous user data of anonymous user data
Palo Alto, USA, and São Paulo, Brazil, May 4, 2022 – Itaú Unibanco, the largest private bank in Latin America, and QC Ware, a leader in quantum software and services, announce the first results of a results of a collaboration that explores the use of quantum computing algorithms in the banking banking sector. The goal of the joint project, which lasted four months, is to investigate whether quantum computing can help with customer retention. The quantum machine learning algorithm proposed by QC Ware improves the performance of the models currently used to predict customer loss.
Throughout this collaboration, the two teams have developed new methods that run on the traditional computers used today and can already improve predictive models, achieving substantial increases in customer retention in back-testing. These algorithms will run even faster on the quantum computers of the future, considering their inherent ability to perform complex linear algebra tasks.
The collaboration has the ongoing goal of gaining knowledge and increasing internal expertise the power of quantum algorithms, preparing Itaú Unibanco for the imminent implementation of quantum computing throughout the financial services sector. An initiative that combines Itaú Unibanco’s experience in banking, machine learning, and quantum computing with QC Ware’s leadership in classical and quantum algorithms.
“Keeping our customers satisfied is one of Itaú Unibanco’s top priorities, and we will continue to We will continue to lead the implementation of innovative technologies”, says Moisés Nascimento, Chief Data Officer at Itaú Unibanco. “We see in quantum computing the potential to greatly improve interactions with customers and we have already benefited from QC Ware’s understanding of existing retention algorithms.” Itaú Unibanco provided QC Ware with years of anonymous user data and approximately 180,000 data points, with the intention of better understanding which customers were likely to leave the bank in the next three months. QC Ware developed quantum methods to train a customer retention model based on a technique known as technique known as determinantal point processes. The quantum methods improved accuracy accuracy and reduced execution times compared to traditional techniques.
The teams at QC Ware and Itaú Unibanco found a way to make a variant of these methods available on today of these methods on current computers, which improved Itaú Unibanco’s model by Unibanco model, increasing the amount of withdrawals captured by 2%, as well as increasing the overall accuracy of the accuracy of the model from 71% to 77.5%. The algorithm can continue to run on traditional computers and is already ready to run on the quantum hardware of the future.
“This project has generated a lot of knowledge for us and a new use of quantum techniques and quantum-inspired determinate sampling techniques to improve machine learning learning,” says Iordanis Kerenidis, head of quantum algorithms at QC Ware. “We are very happy that we have developed powerful quantum methods and found ways to improve performance and efficiency today. We are excited about the prospects of quantum computing in financial services.”