Supratik Paul

Supratik Paul

DPhil Student

My research is on policy search reinforcement learning. I am working on developing algorithms to efficiently learn policies robust to significant rare events - events that can significantly impact the performance of a policy, but have a very low probability of occurrence, e.g. strong wind conditions on an autonomous helicopter. My work involves Gaussian Processes, Bayesian optimisation, Bayesian quadrature, Monte Carlo sampling and model based policy search methods.