Seminar content
It includes an explanation of the main differences between traditional statistical approaches and modern ML/AI algorithms, including examples of their application in the areas of credit, market, and operational risk. Emphasis is placed primarily on the process of back-testing and validation of these models, testing the stability of results, identifying errors, and assessing the models’ ability to respond to changes in data.
A part of the seminar is also devoted to the perspective of regulators (EBA, ECB) on the use and validation of ML/AI-based models, including considerations on model risk, data management, and documentation.