STA 290 Seminar: Michael Berger

seminar thumbnail

Event Date

Location
Mathematical Sciences Building 1147

Speaker: Michael Berger, Head of Insure AI at Munich Re

Title: "Risk Assessment for Machine Learning Models"

Abstract: Machine learning (ML) and Artificial Intelligence (AI) are increasingly used across many industries. However, their uptake in safety-critical areas such as healthcare and finance has been slower due to concerns about model failures and the associated risks. To address these challenges and foster trust in AI, Munich Re, a global leader in reinsurance and insurance, has developed specialized insurance solutions to help companies manage the risks of AI adoption. Central to these solutions is the ability to quantify uncertainty in ML model performance. Leveraging recent advances in statistical methods, including conformal prediction, Munich Re’s researchers have designed flexible and accurate risk assessment tools for black-box ML models. This presentation will explore these methods, their theoretical foundations, and their performance in real-world applications, particularly under both exchangeable and covariate-shift data conditions.

About: Michael Berger is currently the head of Insure AI at Munich Re. He got his PhD in finance from Bundeswehr University Munich.  He has also been named to Capital Magazine’s Top 40 Under 40. 


Seminar Date/Time: Thursday October 3, at 4:10pm

Location: MSB 1147. Refreshments at 3:30pm, MSB Courtyard

Tags