Managing Machine Learning Models in the Financial Industry
Stu Kozola, FRM from MathWorks and Sri Krishnamurthy from QuantUniversity
QuantUniversity partnered with PRMIA to do the first QuantUniversity summer school in Machine Learning and AI in Finance. We had more than 1000 participants from more than 20 countries including India, China, Australia, UK, Turkey, South Africa etc. attend the summer school series.
This year, we offered 3 courses in Data Science, Machine Learning, and Model Risk Management:
- Just Enough Python for Data Science
- Machine Learning and AI for Financial Professionals
- Model Risk Management for Machine Learning Models
In addition, we had 10 lectures from eminent quants, innovators, and thinkers on various topics in AI/ML and Fintech related topics.
In Week 7, we had Stu Kozola, FRM from MathWorks discuss his work on Model Risk Management for AI and Machine Learning. We also had Sri Krishnamurthy from QuantUniversity Discuss Rapid Prototyping Quant Research ML Models using the QUSandbox. Here is a summary of the workshop.
Lecture 1: Model Risk Management for AI and Machine Learning
Artificial intelligence and machine learning are part of today’s modeler’s toolbox for building challenger models and new innovative models that address business needs. However, AI presents new and unique challenges for risk management, particularly for assessing, controlling, and managing model risk for models of limited transparency. Another key consideration is the speed at which these models can be developed, validated, and then deployed into productive use to be competitive adhering to a robust model risk management program. This talk will highlight best practices for integrating AI into model risk practices and showcase examples across the model lifecycle.
Lecture 2: Rapid Prototyping Quant Research ML Models for Algorithmic Auditing using the QuSandbox
Unlike traditional quant models, ML models require constant iteration, tweaking, testing, monitoring, and retuning. Without a rigorous process for facilitating these Agile workflows for machine learning, Quants are destined to be tied up in a brittle process that is not agile nor scalable OR builds models without any process encumbrances incurring major model risks in their workflow. As the scale of ML model adoption increases within the enterprise, a controlled process that enables Quants to be creative and explore tools and datasets of their choice is needed. In this talk, we will illustrate, through a case study on why a Sandbox based approach to building machine learning models is warranted.
Listen to the Podcast here
Slides, demos, and videos at
If you want to try out the demos yourselves on the QuAcademy:
Use ‘QUSUMMERSCHOOL’ as Registration Code