Billy Bourdaniotis
Implementing a numerical scheme for pricing of American options under model uncertainty
The main focus of this project is to derive, implement, and assess a numerical algorithm from backward stochastic differential equations (BSDEs) and cross entropy regularisation to price American options with the assistance of Neural Networks in Python. Research in the past has used machine and reinforcement learning techniques such as Monte-Carlo sampling, and HJB Partial Differential Equations to either price American and Bermudan options, or estimate optimal stopping times. This research aims to combine the use of BDSEs, machine learning and reinforcement learning by accounting for model risk, that is, uncertainty in the underlying asset process, in order to estimate the value of an American option.
Billy Bourdaniotis
The University of New South Wales
Billy Bourdaniotis is a Bachelor of Actuarial Studies/Advanced Mathematics (Honours)
student at the University of New South Wales. He is also an academic tutor and research
assistant at the School of Risk and Actuarial Studies at UNSW and tutor for the School of
Computer Science and Engineering. Through coursework and learning experiences, Billy
discovered financial derivatives such as options, and their underlying complexity, which
became the topic of interest for his SRS project. He is also interested in programming and
deep learning which he has excelled in throughout his coursework. This opportunity
complements his current trajectory to complete an honours year, possibly segueing into
PhD, both focussing on a similar field of mathematics.
You may be interested in
Muhammad Haris Rao
The Representation Theory of Hecke Algebras Through the Knizhnik-Zamolodchikov Functor
