Neil Kichler

PhD Student at RWTH Aachen University

Speaker

I am a PhD student in computer science at RWTH Aachen under the supervision of Prof. Naumann. We are working on Algorithmic Differentiation (AD) in the broadest sense. Recently, my aim is to apply the learnings of AD to the related realm of computing convex and concave relaxations of factorable functions. The application of AD over relaxations leads to relevant subgradients typically used in deterministic global optimization. The flip side, computing relaxations of differentiable programs, opens up new avenues by providing derivative information over entire intervals. Applications thereof are still underexplored. My poster will explore its use for convex relaxations of option prices and the related Greeks. AD provides accurate and efficient ways to compute those Greeks and, in combination with those relaxations, can give insights when parameters are specified by intervals (due to, e.g., model calibration uncertainty). I am also interested in efficient machine learning (ML) with experience in surrogate modeling using differential ML, sparse neural networks, and GPU computing using CUDA.