Topic of the lecture
Path dependent modelling in Finance and Machine Learning
Abstract
We provide a continuous time version of Takens theorem fromdynamical system theory and discuss relations to modelling in mathematical Finance and in Machine Learning. We also discuss two types of architectures for modelling such path dependences from a perspective of invariant theory: transformers and signatures. Finally we introduce a flexible time series model based on signatures.
Where and when
The lecture will take place on August 4, Monday, at 13:00, in room 201 of the Red Building of the Taras Shevchenko National University of Kyiv, 60 Volodymyrska Street.
You can join the lecture online by clicking here
Join Zoom Meeting
https://knu-ua.zoom.us/j/89677817349?pwd=N2lMeHJmcHd4d0p6SWJrMkFYQXhUUT09
Meeting ID: 896 7781 7349
Passcode: 450891
Professor Josef Teichmann from the Federal Institute of Technology Zurich (ETH Zurich) is one of the most famous experts in probability theory today. Among his other results, he is known for his research in machine learning with applications in financial theory – for example, his article on deep hedging has more than 500 citations.
Professor Teichmann has been actively supporting Ukraine and our University since the beginning of the full-scale invasion. In particular, in 2022, he gave a course of lectures at the Faculty of Mechanics and Mathematics on the application of machine learning in finance, and in 2023, he helped to attract professors from other Western universities to teach a standard course on machine learning.