Matt Moores

Lecturer in Statistical Science

University of Wollongong

Biography

Matt Moores is a lecturer in the School of Mathematics and Applied Statistics (SMAS) at the University of Wollongong. He is also affiliated with the National Institute for Applied Statistics Research Australia (NIASRA). His research interests include Bayesian inference and scalable computation for emerging applications in hyperspectral imaging, such as Raman spectroscopy and satellite remote sensing.

Prior to joining Wollongong, Matt was a Research Fellow in the Department of Statistics at the University of Warwick. There, he worked on the EPSRC-funded projects i-like: “Intractable Likelihood – New Challenges from Modern Applications” and “in situ Nanoparticle Assemblies for Medical Diagnostics and Therapy.” Matt completed his PhD in 2015, under the supervision of Prof. Kerrie Mengersen and Dr Fiona Harden at Queensland University of Technology (QUT), Brisbane, Australia. His thesis topic was “Bayesian computational methods for spatial analysis of images.” Before commencing his PhD, Matt was involved in the Visible Cell project at the Institute for Molecular Bioscience (IMB), University of Queensland. He also has extensive industry experience, having worked in R&D for international companies such as Oracle, Digital (DEC), and Compaq. Matt is the co-inventor of 3 U.S. patents in the field of information retrieval.

Interests

Bayesian Inference
Computational Statistics
Spectroscopy
Image Analysis
R

Education

PhD in Statistics, 2015

Queensland University of Technology
MMathSci, 2008

Queensland University of Technology
BInfTech (Computing Science), 1996

Queensland University of Technology

Selected Publications

Scalable Bayesian Inference for the Inverse Temperature of a Hidden Potts Model

This paper introduces the parametric functional approximate Bayesian (PFAB) algorithm to accelerate computation for the intractable likelihood of the Potts model. We demonstrate this method using synthetic data as well as remotely-sensed imagery from the Landsat-8 satellite. We achieve up to a hundredfold improvement in the elapsed runtime, compared to the exchange algorithm or ABC. An open source implementation of our algorithm is available in the R package bayesImageS.

MT Moores, GK Nicholls, AN Pettitt, K Mengersen

arXiv:1503.08066v3 [stat.CO], 2018

Preprint Code Project Slides Video

Bayesian modelling and quantification of Raman spectroscopy

We introduce a sequential Monte Carlo (SMC) algorithm to separate an observed spectrum into a series of peaks plus a smoothly-varying baseline, corrupted by additive white noise. The peaks are modelled as Lorentzian, Gaussian, or pseudo-Voigt functions, while the baseline is estimated using a penalised cubic spline. An open source implementation of our algorithm is available in the R package serrsBayes.

MT Moores, K Gracie, J Carson, K Faulds, D Graham, M Girolami

arXiv:1604.07299v2 [stat.AP], 2018

Preprint Code Dataset Project Slides

Recent Publications

MT Moores, GK Nicholls, AN Pettitt, K Mengersen. Scalable Bayesian Inference for the Inverse Temperature of a Hidden Potts Model. arXiv:1503.08066v3 [stat.CO], 2018.

Preprint Code Project Slides Video

CC Drovandi, MT Moores, RJ Boys. Accelerating pseudo-marginal MCMC using Gaussian processes. CSDA 118: 1-17, 2018.

Preprint Code Project Slides Source Document

MT Moores, K Gracie, J Carson, K Faulds, D Graham, M Girolami. Bayesian modelling and quantification of Raman spectroscopy. arXiv:1604.07299v2 [stat.AP], 2018.

Preprint Code Dataset Project Slides

K Gracie, MT Moores, WE Smith, K Harding, M Girolami, D Graham, K Faulds. Preferential Attachment of Specific Fluorescent Dyes and Dye Labeled DNA Sequences in a Surface Enhanced Raman Scattering Multiplex. Anal. Chem. 88(2): 1147–1153, 2016.

Code Dataset Project Source Document

MT Moores, CE Hargrave, T Deegan, M Poulsen, F Harden, K Mengersen. An external field prior for the hidden Potts model with application to cone-beam computed tomography. CSDA 86: 27-41, 2015.

Preprint Code Project Source Document

MT Moores, CC Drovandi, K Mengersen, CP Robert. Pre-processing for approximate Bayesian computation in image analysis. Stat. Comput. 25(1): 23-33, 2015.

Preprint Code Project Source Document

Recent & Upcoming Talks

Computational Statistics and Molecular Simulation: A Practical Cross-Fertilization

Nov 12, 2018 12:00 AM

CMO-BIRS Workshop 18w5023

Code Project

bayesImageS: an R package for Bayesian image analysis

Jul 11, 2018 3:30 PM

useR! 2018

Slides Video Code Project

Sequential Monte Carlo for Multivariate Calibration in Chemometrics

Jun 25, 2018 6:00 PM

ISBA World Meeting 2018

Slides Code Project

Projects

bayesImageS

Bayesian algorithms for image segmentation.

serrsBayes

Bayesian computation and inference for spectroscopy.

Research Supervision

I am available to supervise Honours, Masters, and PhD students in the School of Mathematics and Applied Statistics at Wollongong. If the above projects in spectroscopy and MCMC appeal to you, please get in touch with me via email.

Current Masters Students

Spring 2018 / Autumn 2019:

Lu Wang, “Experimental design for Multiplex Spectroscopy”

Past Masters Students

Jianyin Peng (MSc with distinction, University of Warwick, 2018) “A New Errors-in-variables Model and Bayesian Computation for Raman Spectra”

Teaching

I am a lecturer for the following courses in the Spring 2018 session at Wollongong:

STAT101: Introduction to Statistics (coordinator: Dr Carole Birrell)
PSYC354: Design and Analysis (coordinator: Dr Sebastien Miellet)

Contact

mmoores@uow.edu.au
02 4298 1358
mt_moores
39C.190, University of Wollongong, Northfields Ave, NSW 2522, Australia
Monday 14:00 to 16:00 & Wednesday 14:00 to 16:00