Jason Michael Rader
University of Oxford
Andrew Wiles Building
Radcliffe Observatory Quarter
- C6.1: Numerical Linear Algebra (TA)
- B8.1: Probability, Measure and Martingales (TA)
- MSc MCF Deep Learning (TA)
Hello! I'm Jason, a first year student at the CDT in Mathematics of Random Systems working with Terry Lyons and Ben Hambly. I am also a postgraduate member of DataSig, a joint research group between Oxford, Imperial, UCL, and the Alan Turing Institute focused on understanding complex, multimodal data streams.
I'm generally interested in blending data driven modeling and mathematical modeling to solve problems with partially known structure.
Concretely, I have recently been working on higher-order, memory efficient numerical methods for neural differential equations. Neural differential equations are models which use deep learning to (among many other things) capture the unknown portion of vector fields arising in traditional differential equation modeling. I am also currently using deep learning to price complex interest rate derivatives by extending simple interest rate models to arbitrage-free models capable of capturing more market complexity.
If this sounds interesting, or if you have such a problem, please don't hesitate to reach out!