PREMIUM CONTENT
This is premium content, available to Connect Plus users only
Unlock this content and more with Connect Plus membership.
Join a community of professionals and get:
Join a community of professionals and get:
15% discount
on all CeFPro events.
on all CeFPro events.
Post-event access:
unlock speaker decks and audience polls.
unlock speaker decks and audience polls.
Instant insights:
Full library access the moment you sign up.
Full library access the moment you sign up.
Article
Machine Learning Meets Market Turbulence as Risk Control Tightens
As machine learning accelerates across financial services, its weakest point remains financial market time series, where noise, regime shifts and short histories undermine model stability. Peter Quell argues that banks must fundamentally rethink validation, data strategy and model classification frameworks to ensure ML behaves reliably under stress. The challenge is not enthusiasm for AI, but disciplined governance that keeps pace with markets that refuse to stand still.
Dec 17, 2025
Peter Quell, Head of Portfolio Analytics for Market and Credit Risk, DZ Bank AG
Tags:
AI and Technology (including Fintech)
The views and opinions expressed in this content are those of the thought leader as an individual and are not attributed to CeFPro or any other organization
• ML struggles with noisy, non-stationary financial time series
• Overfitting remains the central threat to market-risk ML models
• Banks face major gaps moving ML prototypes into production
• Traditional model-risk frameworks are not built for ML complexity
• Data quality, structure and volume introduce new failure modes
• Continuous recalibration demands new governance and oversight
• Overfitting remains the central threat to market-risk ML models
• Banks face major gaps moving ML prototypes into production
• Traditional model-risk frameworks are not built for ML complexity
• Data quality, structure and volume introduce new failure modes
• Continuous recalibration demands new governance and oversight
Log in to continue or register for free
WHAT'S INCLUDED:
Unlimited access to peer-contribution articles and insights
Global research and market intelligence reports
Discover Connect Magazine, a monthly publication
Panel discussion and presentation recordings
Sign in to view comments
Related insights —