A Deep Learning Scheme for Solving Fully Nonlinear Partial Differential Equations
Watch a 50-minute technical presentation from the SIAM Activity Group on FME Virtual Talk Series exploring convergence analysis of deep learning algorithms for solving fully nonlinear partial differential equations. Dive into the mathematical framework showing how finite difference approximations to the deep learning loss function enable convergence to unique viscosity solutions. Examine practical applications through case studies of finite horizon optimal investment problems involving proportional transaction costs in both single and multi-asset scenarios. Learn advanced techniques at the intersection of deep learning, numerical analysis, and financial mathematics from expert researcher Maxim Bichuch of SUNY Buffalo as he demonstrates novel approaches for solving complex PDEs relevant to quantitative finance and computational science.