I'm a fourth-year PhD student at the Univeristy of Bristol's ESPRC Centre for Doctoral Training in Computational Statistics and Data Science working on data compression.
Check out my personal website for additional details about me!
Conditional Distribution Compression via the Kernel Conditional Mean Embedding - Published at NeurIPS 2025, this work extends distribution compression to act on joint and conditional distributions. Distribution compression is concerned with reducing the size of datasets while maintaining downstream performance, in our case we consider various regression, classification and uncertainty quantification tasks.
Bilateral Distribution Compression: Reducing Both Data Size and Dimensionality - Under review at ICML 2026, this work extends distribution compression to also consider dimension reduction. We reduce both the number of observations adn the number of dimensions of a dataset simultaneously while maintaing or improving downstream task performance for datasets which exhibit low-dimensional manifold structure.
Kernel Herding for distribution compression - https://github.com/dominicjbroadbent/kherd
Spectral conditional density estimation in high dimensions - https://github.com/dominicjbroadbent/SpectralCDE
Company pivot / rebrand detector - https://github.com/dominicjbroadbent/pr_detector
Predicting film viewer ratings - https://github.com/dominicjbroadbent/film_ratings
Local Regression - https://github.com/dominicjbroadbent/loess
Univariate Kernel Density Estimation - https://github.com/dominicjbroadbent/kde
Module for Gaussian Process classification - https://github.com/Tennessee-Wallaceh/gproc
Developing a Computer Algebra System (CAS) - https://drive.google.com/file/d/1Pmz9MtWIYuMLyR8UbUFo2uV7aZK6c2zj/view?usp=sharing
Modelling a Measles Epidemic - https://drive.google.com/file/d/1PiZONJ8Mxpj-EOTMQR7U_rgqJN-SMF4F/view?usp=sharing
Website - https://dominicjbroadbent.github.io/
University Profile - https://compass.blogs.bristol.ac.uk/students/dominic-broadbent/