- The company is advancing its quantum computing strategy and has developed a machine learning algorithm to classify customers according to their credit risk.
- With this project, CaixaBank has become the first institution in Spain and one of the first in the world to incorporate quantum computing into its services.
After successfully performing the first real tests of quantum computing to study the applications of this technology in financial services, CaixaBank has taken a step further and developed the first machine learning algorithm to classify risks in Spanish banking leveraging quantum computing.
Chaired by Jordi Gual and with CEO Gonzalo Gortázar, the bank becomes the first Spanish financial institution to apply a hybrid computing framework — which combines quantum computing and conventional computing in different phases of the calculation process — to classify credit risk profiles. To do this, CaixaBank used a public data set corresponding to 1,000 artificial users, with a similar profile to existing customers, but with information configured specifically for the test.
With this project, the institution is making improvements in risk scenario simulations and machine learning, underpinning increasingly complex algorithms which require large quantities of data to learn, whilst also progressing its analysis of quantum computing applications. The results of this test, which demonstrates that hybrid computing can achieve results comparable to those offered by the conventional solution in less time, will be published in more detail in specialist channels so that the conclusions are available to the community.
Quantum computers are based on the properties of superconductors, which integrate their process units, known as qubits, instead of the classical [binary] bits. Due to these properties, they have the capacity to process a multitude of variables and scenarios simultaneously, achieving a computing capacity that grows exponentially with the number of qubits.
Hybrid computing uses this exponential computing advantage to perform complex calculations of parameters optimizing machine learning algorithms and combines them with classical computing methods to make the most out of both systems. With the application of hybrid algorithms (quantum and classical) in risk analysis, the institution can reach the same conclusions as the classical method in much less time.
Prior to this solution, CaixaBank developed a project to carry out risk assessment simulations for financial assets using quantum computing. In this field, the bank implemented a quantum algorithm capable of assessing the financial risk of two portfolios created specifically for the project based on real data, one consisting of mortgages and the other, treasury bills.