Mathematical Sciences Department
Joint Colloquium and Financial Math Seminar
Speaker: Michael Ludkovski, University of California, Santa Barbara
Friday, December 1, 2023
2:00 pm - 2:50 pm
Salisbury Labs 402
Title: Statistical Surrogates in Quantitative Finance: from Learning the Greeks to Adaptive Robust Stochastic Control
Abstract: I will start by surveying the use of statistical surrogates, also known as functional approximators, for learning input-output relationships in quantitative finance tasks. I will review the main ingredients of surrogate construction: experimental design, loss function selection and training method, and then summarize the two major surrogate classes of Gaussian process regression and (deep) neural networks. In the second half of the talk, I will present case studies on learning option sensitivities (Saporito and Ludkovski, AMF 2022), pricing Bermudan options (Ludkovski, JCF, 2023) and adaptive robust utility maximization (Chen and Ludkovski, SIFIN 2021).