The use of machine learning (“ML”) models in banks has raised considerable interest and sparked a lively debate. In November 2021, the European Banking Authority (“EBA”) has published a discussion paper outlining principle-based recommendations on their use for regulatory purposes. ML models can provide significant benefits to banks, especially with dealing with unstructured data and complex relationships. Yet, concerns have been raised, above all about data quality and model interpretability, that need to be addressed carefully if such models are to be successfully deployed.
Program
9:45 | Introductory remarks, Andrea Resti – Bocconi University and advisor to the European Parliament on banking supervision
10:00 | “ML and credit risk: from managerial models to validated IRB modules ” Dario Cavarero, Head of Sme Models Office, Intesa Sanpaolo
10:20 | “ML for early warning models with high-frequency account transaction data” Virginia Tirri, Head of Performance Management Bonis, CLO-GC, BancoBPM
10:40 | “Credit Risk and ML models: challenges and opportunities for UniCredit Group” Aurelio Maccario, Head of Group Credit Risk, UniCredit Group
11:00 | “The contribution of ML to model validation and Level 2 checks” Rita Gnutti, Executive Director, Internal Validation and Controls, Intesa Sanpaolo
11:20 | “A ML-based model supporting credit origination policies” Daniele Vergari, Director, Risk Analytics - Transformation Services, CRIF
11:40 |“ML models and banks: views from an ECB expert” Hendrik Brakemeier, AI & Machine Learning Application Development Expert, European Central Bank
12:00 | “EBA work on ML models in credit risk: early results and possible future developments” Roberta De Filippis, policy expert in prudential regulation, European Banking Authority