Speaker |
Louis Sharrock |
|---|---|
Affiliation |
University College London |
Date |
Wednesday, 10 December 2025 |
Time |
12:30-13:30 |
Location |
Ground floor lecture theatre, Sainsbury Wellcome Center, 25 Howland St, W1T 4JG |
Link |
https://ucl.zoom.us/j/99748820264 |
Event series |
Jump Trading/ELLIS CSML Seminar Series |
Abstract |
The task of sampling from a target probability distribution known only up to a normalisation constant is of fundamental importance to computational statistics and machine learning. There are various popular approaches to this task, including Markov chain Monte Carlo (MCMC) and variational inference (VI). Unfortunately, such methods invariably depend on hyper-parameters such as the step size, which must be carefully tuned by practitioners in order to ensure convergence to the target distribution at a suitable rate. In this talk, we introduce a suite of new sampling algorithms which are entirely step-size free. Our approach leverages the perspective of sampling as an optimisation problem over the space of probability measures, and existing ideas from convex optimisation. We discuss how to establish the convergence of our algorithms under assumptions on the target distribution. We then illustrate the performance of our approach on a range of numerical examples, demonstrating comparable performance to existing algorithms, but with no need to tune a step size. |
Biography |
I am a Lecturer (Assistant Professor) in the Department of Statistical Science at University College London. I was previously a Senior Research Associate with Prof. Chris Nemeth at Lancaster University, and a Data Science Heilbronn Research Fellow at the University of Bristol. I obtained my PhD in the Department of Mathematics at Imperial College London, supervised by Dr. Nikolas Kantas. I also hold an MRes in Mathematics and an MSc in Statistics from Imperial College London, and an MA in Mathematics from the University of Cambridge. |