TopQuants

Jörg Kienitz

mrig and University of Cape Town

GenAI - A Gaussian Perspective

For synthetic market data generation we consider Gaussian methods. Our research is based on applying Gaussian Mixture Models and Control Variates when applicable. The idea is to approximate the distribution of the observed variables or invariants like returns. We illustrate the method using time series imputation, backfilling, yield curve and volatility surface simulation. We stress that the method is applicable with a reasonable amount of data. In the talk we discuss the methodology, examples, and theoretical aspects as no-arbitrage.

Jörg Kienitz is currently Director – Quantitative Methods at mrig, a Frankfurt based consultancy company, Adjunct Associate Professor at the African Institute of Financial Markets and Risk Management (AIFMRM) at the University of Cape Town. Furthermore, he consults to Standard Bank Insurance and Asset Management (LifFin). He works on all aspect of modelling the financial markets including derivatives pricing, real world modelling, risk management, model validation and applying machine learning methods.

Prior to joining mrig Jörg hold positions at LSEG, acadia, Deloitte and Postbank/Deutsche Bank as Head of Quant. The main focus of Jörg’s research is on Quantiative Finance and Machine Learning methods. He is a regular speaker at major conferences including WBS Quant Conference, Quant Minds or RISK and is a published author with Wiley and Palgrave/Springer. His research appeared in major journals including Quantitative Finance, Journal of Computational Finance, RISK, or Wilmott.