This course aims to provide an overview and understanding of common Machine Learning (ML) techniques, to understand the opportunities and limitations of these techniques and to be able to interact with the experts from ESAs, academia, and industry. Participants will work hands-on with ML/AI methods in Python and demystify the black boxes to continue learning by themselves.
The training week is structured to encompass a series of lectures and complementary tutorials focused on ML/AI techniques, ensuring that participants not only grasp the theoretical foundations of ML/AI techniques but also acquire practical skills through guided tutorials. Links to key regulatory frameworks are made in this course. This part of the training will delve into the implications of these advancements within the context of the financial industry and supervisory practices. This training week is designed for participants who already have some exposure to coding and are looking to enhance their practical knowledge of ML/AI techniques.
Introduction to statistics for AI and ML
Predictive models and linear regression
Classification models and logistic regression
Feature selection and regularization
Unsupervised learning and clustering
Finding clusters and neighbors
Natural Language Processing (NLP)
Python applications of the theoretical sessions
High-level overview of related regulations: new EU rules on AI – proposal for a regulation on AI : regulatory and supervisor implications for the financial sector, including requirements on explainability
NCAs use case of SupTech applications
Supervisory dialogue with the industry on future market developments: Buy vs. Build SupTech solutions
Essential: Quantitative background – Economics, Finance or Computer Science or equivalent degree
Essential: Prior coding experience, preferably with Python
Recommended: Previous exposure to the design and/or use of SupTech application