How Bankia applies and understands explainable AI
Bankia is one of the Spanish banking institutions that is most committed to the use of Artificial Intelligence (AI) in its processes as a tool to offer its customers the products and services that best suit their needs. So much so that as a result of its 2018-2020 strategic plan, the Centre of Excellence in Artificial Intelligence was created to continue implementing the use of AI in the organization.
Jaime Zamora, member of the Center of Excellence in AI of Bankia, explains first hand some of the most relevant projects they are carrying out, and also how they should address the challenges and problems that sometimes arise from the use of this technology in a market as complex, regulated and strategic as the banking one.
“An after-control processing flow allows us to ensure correct decision making of our AI models”.
“Our platforms do not make automated decisions based exclusively on the results offered by our artificial intelligence models, but there is a subsequent control processing flow that allows us to ensure the correct decision making in each case,” says Zamora.
“We do not conceive of data science in the simple use of the multitude of existing standard Machine Learning programming libraries, but in the mathematical understanding of our models”.
Avoiding biases in the decisions that are made is one of the main objectives, and this necessarily involves understanding and being able to explain the nature of each Machine Learning model, of each neural network, and thus being able to “adjust and calibrate our models correctly and at all times,” concludes Zamora, who adds: “we do not conceive of data science in the simple use of the multitude of standard programming libraries that already exist for Machine Learning, but in the mathematical understanding of our models. There is no Science if there is no knowledge, we are a centre of excellence”.