About
I am a Research Fellow at the University of Technology Sydney and a Visiting Scientist at CSIRO’s Data61. My research focuses on developing new techniques for machine learning on complex data, with an emphasis on approaches that are statistically principled, computationally scalable, and operationally trustworthy.
The main directions of my research fall under a few broad areas of machine learning:
- Causal discovery
Uncovering the underlying mechanisms that generate observed data, separating true cause-and-effect relationships from spurious associations to enable deeper scientific understanding. - High-dimensional learning
Extracting reliable structure and signal from datasets with thousands or millions of measured variables, where classical statistical principles break down and new techniques are needed. - Bayesian inference
Using probabilistic reasoning to integrate prior knowledge with new evidence, yielding transparent measures of uncertainty that guide reliable decision-making.
Before my current role, I was a Research Associate at the University of New South Wales, following a PhD at Monash University. Earlier in my career, I worked at KPMG, advising financial institutions on quantitative risk management. I completed my undergraduate studies at the University of Sydney.
My curriculum vitae is available here.
Research
ProDAG: Projected variational inference for directed acyclic graphs
Ryan Thompson, Edwin Bonilla, and Robert Kohn
Advances in Neural Information Processing Systems, 2025
Preprint - Code
Scalable subset selection in linear mixed models
Ryan Thompson, Matt Wand, and Joanna Wang
arXiv, 2025
Preprint - Code
Semi-supervised Gaussian mixture modelling with a missing data mechanism in R
Ziyang Lyu, Daniel Ahfock, Ryan Thompson, and Geoffrey McLachlan
Australian and New Zealand Journal of Statistics, 2024
Publication - Preprint - Code
Contextual directed acyclic graphs
Ryan Thompson, Edwin Bonilla, and Robert Kohn
International Conference on Artificial Intelligence and Statistics, 2024
Publication - Preprint - Code
Familial inference: Tests for hypotheses on a family of centres
Ryan Thompson, Catherine Forbes, Steven MacEachern, and Mario Peruggia
Biometrika, 2024
Publication - Preprint - Code
Flexible global forecast combinations
Ryan Thompson, Yilin Qian, and Andrey Vasnev
Omega, 2024
Publication - Preprint - Code
Group selection and shrinkage: Structured sparsity for semiparametric additive models
Ryan Thompson and Farshid Vahid
Journal of Computational and Graphical Statistics, 2024
Publication - Preprint - Code
The contextual lasso: Sparse linear models via deep neural networks
Ryan Thompson, Amir Dezfouli, and Robert Kohn
Advances in Neural Information Processing Systems, 2023
Publication - Preprint - Code
Robust subset selection
Ryan Thompson
Computational Statistics and Data Analysis, 2022
Publication - Preprint - Code
Optimal selection of expert forecasts with integer programming
Dmytro Matsypura, Ryan Thompson, and Andrey Vasnev
Omega, 2018
Publication - Preprint