Matthew Sutton

I'm a Lecturer in the school of mathematical sciences at QUT. My research interests span broadly across Monte Carlo methods, high-dimensional statistics, and Bayesian methodology. I've worked on penalised regression methods, dimension reduction, partial least squares, Bayesian inference, Markov chain Monte Carlo (MCMC). I am currently working on the Models and Algorithms research program for the center for data science.

Continuous-time Monte Carlo

My current research directions are in developing fast Monte Carlo methods. I've been recently working on a class of algorithms known as Piecewise Deterministic Markov Processes (PDMPs). These produce a Markov process instead of the standard Markov chain used in classical MCMC and have a number of very interesting properties. Notabley PDMPs:

1. Scale computation for big data by sub-sampling the data at each iteration and remaining asymptotically exact.
2. Are non-reversible methods that leverage gradient information to enhance mixing properties.
3. Maintain non-reversiblity even in regions of space where the gradient is undefined.
4. Can be constructed as continuous-time limits of existing MCMC methods.

Below is an illustration of a simple PDMP known as the Zig-Zag process:

In-Press

• Corenflos, A., Sutton, M., & Chopin, N. Debiasing Piecewise Deterministic Markov Process samplers using couplings. In arXiv [stat.CO].

Publications

• Sutton, M., Salomone, R., Chevallier, A., & Fearnhead, P. (2022). Continuously tempered PDMP samplers. Advances in Neural Information Processing Systems.
• Chevallier, A., Fearnhead, P., & Sutton, M. (2023). Reversible Jump PDMP Samplers for Variable Selection. Journal of the American Statistical Association, 118(544), 2915–2927.
• Sutton, M., & Fearnhead, P. (2023). Concave-Convex PDMP-based Sampling. Journal of Computational and Graphical Statistics: A Joint Publication of American Statistical Association, Institute of Mathematical Statistics, Interface Foundation of North America, 32(4), 1425–1435.
• Davies, L., Salomone, R., Sutton, M., & Drovandi, C. (2022). Transport Reversible Jump Proposals. International Conference on Artificial Intelligence and Statistics, vol. 206, pp. 6839–6852.
• Sutton, M., Sugier, P.-E., Truong, T., & Liquet, B. (2022). Leveraging pleiotropic association using sparse group variable selection in genomics data. BMC Medical Research Methodology, 22(1), 9.
• de Micheaux, P. L., Liquet, B., & Sutton, M. (2019). PLS for Big Data: A unified parallel algorithm for regularised group PLS. Statistics Surveys, 13(none), 119–149.
• Sutton, M., Thiébaut, R., & Liquet, B. (2018). Sparse partial least squares with group and subgroup structure. Statistics in Medicine, 37(23), 3338–3356.
• Liquet, B., Mengersen, K., Pettitt, A. N., & Sutton, M. (2017). Bayesian Variable Selection Regression of Multivariate Responses for Group Data. Bayesian Analysis, 12(4), 1039–1067.
• Billington, E.J., Khodkar, A, Petrusma, D and Sutton, M (2014) Lambda-fold theta graphs: Metamorphosis into 6-cycles. AKCE International Journal of Graphs and Combinatorics, 11 1: 81-94.