Tuesday, 4 August 2026, 2pm to 3pm
Abstract: Bayesian statistics is now a standard tool in statistician and machine learner’s toolkit. However, the practical application of Bayesian models and methods can still present challenges and, in turn, reveal interesting new research directions.
In this presentation, I will discuss some of our attempts to create software products that are founded on Bayesian methods. The first are interactive data platforms, the Australian Cancer Atlas and AusEnHealth, which have in turn inspired similar products. The research interests inspired by this work include federated learning, outlier detection and the development of vulnerability indices. The second are interactive Bayesian network products, which have inspired careful consideration of priors for Bayesian models and the development of LLMs to enhance user engagement.
The work that I will discuss has been led by postgraduate students and postdoctoral researchers in the Centre for Data Science at QUT, Australia. I will acknowledge them individually in my presentation.
Selected References
Australian Cancer Atlas. https://atlas.cancer.org.au/
Bon J, A Bretherton, K Buchhorn, S Cramb, C Drovandi, C Hassan, A Jenner, H Mayfield, J. McGree, K Mengersen, A Price, R Salomone, E Santos-Fernandez, J Vercelloni & X Wang (2023) Being Bayesian in the 2020s: opportunities and challenges in the practice of modern applied Bayesian statistics. Philosophical Transactions. Series A.
Bretherton A, Bon J, Warne D, Mengersen K, Drovandi C, (2026) A Principled Approach to Bayesian Transfer Learning, Bayesian Analysis. To appear.
Davoudabi M, et al. (2026) Federated learning moment propagation. In preparation.
Hassan C (2024) Structured Models and Algorithms for Sensitive Data. PhD Thesis, Queensland University of Technology, Australia.
Johnson S, J Toevs, K Mengersen (2026) Using AI to make Complex Systems less complex. Under review.
Price A, M Rigby, P Fiévez, K Mengersen (2025) A spatial vulnerability index for environmental health. Ecological Indicators.
Santos-Fernandez E, S Denman, K Mengersen (2025) New Bayesian and deep learning spatio-temporal models can reveal anomalies in sensor data more effectively. Water Research.
Tuyl F, R Gerlach, K Mengersen (2024) On the Certainty of an Inductive Inference: The Binomial Case. Statistical Science.
Speaker(s): Professor Kerrie Mengersen (Queensland University of Technology)
Department: Statistics (Department)
Organiser: Professor Geoff Nicholls
Organiser email: geoff.nicholls@stats.ox.ac.uk
Host: Professor Geoff Nicholls