CaixaBank develops the first risk classification model for Spanish banks using quantum computing

CaixaBank develops a machine learning algorithm for the classification of clients according to credit risk. „With the completion of these projects, CaixaBank became the first entity in Spain and one of the first in the world to incorporate quantum computing into its innovation activity.”, according to the press release.

CaixaBank advances in its strategy to prepare for the arrival of quantum computing. After successfully carrying out the first real tests of quantum computing to study applications of this technology to the financial business, the entity has gone one step further and has developed the first machine learning algorithm for the risk classification of Spanish banks using quantum computing.

The entity, chaired by Jordi Gual and whose CEO is Gonzalo Gortázar, thus becomes the first Spanish entity to apply a hybrid computing scheme –which combines quantum computing and conventional computing during different stages of the process during the calculation process- to classify credit risk profiles. To do this, CaixaBank has used a public data set (public data set ), corresponding to 1,000 alleged users, with a profile very similar to real customers, but with fully figurative information for the test.

With this project, the entity improves in the simulation of risk scenarios and machine learning , where algorithms are increasingly complex and require large amounts of data to learn, while progressing in its analysis of the quantum computing applications. The results of this test, which shows that hybrid computing enables results comparable to those offered by the classic solution to be achieved in less time, will be published in greater detail on specialized channels so that the conclusions are available to the community.

Quantum computers rely on the properties of superconductors, which integrate their process units, qubits, rather than classic bits. Thanks to these properties, it has the ability to process a multitude of variables and states at the same time, achieving a computational capacity that grows exponentially with the number of qubits.

Hybrid computing takes advantage of this exponential computing advantage for the complex calculation of machine learning algorithm optimization parameters and combines them with classical computing methods, taking advantage of the best of both worlds. 

„With the application of hybrid algorithms (quantum and classical) in risk analysis, the same conclusions can be reached as with the classical method in much less time.”, the bank said.

First entity in Spain to work with quantum computing

Prior to this development, CaixaBank launched a project to carry out simulations to assess the risk of financial assets. In this field, the entity implemented a quantum algorithm capable of evaluating the financial risk of two portfolios specifically created by the project based on real data, one for mortgages and the other for Treasury bonds.

With the completion of these projects, CaixaBank became the first entity in Spain and one of the first in the world to incorporate quantum computing into its innovation activity.

This exercise has enabled the entity to train in the deployment of quantum versions of classical algorithms and to validate the convergence of the quantum solution. It is essential for CaixaBank to invest in exploring the use potential that quantum computing has for the various areas of the financial business, although the first commercial applications may take time to arrive. CaixaBank will continue to explore the use cases and the potential disruptor of quantum computing in the financial world.

###

This interest in digital transformation, constant in the history of CaixaBank, has allowed the entity to have the largest digital customer base in Spain (6.5 million). In addition, CaixaBank has developed projects that have marked technological milestones in the sector, such as the creation of the first ATMs that allow withdrawals by facial recognition and without having to enter the PIN.