Prabh Bhambra
Join the CDT as I thought it would be a great opportunity to return to my passion for physics, and build up the machine learning skills I picked up during my time in banking.
5 January 2020
Project title: Using AI to explain astronomy
Research Group: Astrophysics
Supervisor(s): Prof Benjamin Joachimi & Prof Ofer Lahav
Introduction:
I graduated with an MSci in Physics from UCL with my thesis being on improving the efficiency of electromagnetic accelerators. After graduating, I worked as a quantitative analyst in equity trading for four years. I used stochastic modelling to price exotic equity trades, and machine learning techniques to model market behaviour. I left banking to join the CDT as I thought it would be a great opportunity to return to my passion of physics, and build up the machine learning skills I picked up during my time in banking.
Project description:
UCL has been heavily involved in large galaxy surveys: DES & KiDS (both with their observations complete, many results already published, and new analyses underway), DESI (had its first light in October 2019; survey to start in 2020), Euclid & LSST (both to start surveys in 2022). While there is a well-defined Bayesian model-dependent methodology to derive cosmological parameters from the data, the cosmology community is only at the beginning of getting AI algorithms to (i) explain and interpret what the algorithms are actually doing; (ii) incorporating known base-line Physics in the algorithms; and (iii) discovering new Physics (or new systematics) from the data.
First year group project: KageNova
Placement: Sports Wellbeing Analytics