Special Data Science Seminar - Maxim Bichuch - Deep PDE Solution of BSDE

Description

Speaker: Maxim Bichuch, Johns Hopkins University

Title: Deep PDE Solution of BSDE

Abstract: This talk will investigate the numerical convergence of a solution to a Backward Stochastic Differential Equation (BSDE) using a deep learning method. We transform the BSDE problem to a problem with a coupled Partial Differential Equations (PDEs), and find the relationship between the solutions to the two problems. We find sufficient conditions for the solution of the coupled PDE to be classical, and use a deep Galerkin method to obtain an approximation to the solution of the BSDE. We illustrate this with numerical examples.



Location: Virtual Zoom meeting
https://mcmaster.zoom.us/j/94953730428?pwd=OVdhUWs0K3paQUJobjQ0bkJlbnJEQT09

Meeting ID: 949 5373 0428
Passcode: 126157
Go Back
McMaster University - Faculty of Science | Math & Stats