Moran Medal

Associate Professor David Frazier, Monash University

A web of complex models underpins modern life. Models are used to predict traffic patterns, help control invasive pest populations and mitigate the spread of disease. These models are driven by unknown quantities, and so statistical inference is used to quantify and understand these unknowns, with Bayesian statistical inference methods often applied in such settings due to their interpretability. However, in many cases the underlying models and data are so complex as to render standard Bayesian methods intractable. In such cases, the best we can hope to do is perform statistical inference using ‘approximate’ Bayesian methods, which seek to deliver tractable Bayesian inferences in challenging modeling settings. Much of Associate Professor David Frazier’s research has focused on establishing the statistical behaviour of approximate Bayesian methods in a wide variety of contexts, including approximate Bayesian computation, Bayesian synthetic likelihood, and variational Bayes methods. The overarching goal of his work is to ensure practitioners can reliably apply these approximation methods to derive meaningful inferences, make reliable decisions and obtain reproducible results.

Dr Rachel Wang, University of Sydney

Working at the interface of theoretical statistics, computational statistics and data-driven applied fields, Dr Rachel Wang has pursued a diverse research trajectory emphasising both rigorous theoretical development and practical relevance to interdisciplinary scientific problems. She has made contributions to statistical inference problems in network models, enabling model selection and parameter tuning to be performed with provable guarantees. Her theoretical work on local convergence issues in variational approximation and scalable MCMC has led to a deeper understanding of how algorithms navigate a high dimensional, non-convex landscape, addressing a prevalent problem in all large-scale machine learning tasks. Leveraging her expertise in theory and computation, she has developed novel statistical and computational tools for extracting new biological knowledge from genomics data, seeking to improve our understanding of gene regulatory mechanisms and the inner workings of cells.

Professor Christopher Drovandi, Queensland University of Technology

Almost every field of science requires sophisticated data analysis, and this in turn requires increasingly sophisticated methods for intelligent data collection and efficient computation. Professor Drovandi's research contributes substantively to both of these areas. He has created new methods for optimal design of experiments that facilitate more cost-effective, data-substantiated decision-making. His innovative research into synthetic likelihood estimation have freed traditional constraints of likelihood-based statistical modelling and computation. His application of these methods to diverse problems in computational biology and exercise science have generated new insights for scientists and managers in these fields.

Dr Janice Scealy, Australian National University

Dr Scealy’s research focuses predominantly on developing new statistical analysis methods for data with complicated constraints including compositional data (vectors of proportions which sum to one), spherical data, directional data and manifold-valued data defined on more general curved surfaces. Her work has led to important new insights in a diverse range of applications. Her new flexible compositional model was applied to predict the proportions of total weekly expenditure on food and housing costs in Australia. Janice used a manifold data transformation to help identify geochemical processes acting on the surface of the Australian crust. She has developed multiple new statistical techniques for analysing noisy paleomagnetic datasets and her methods have led to improvements in uncertainty measurements of Earth’s magnetic field.

Dr Kim-Anh Lê Cao, University of Melbourne

The main focus of Dr Lê Cao’s research is to develop statistical and computational methods that are applicable to high-throughput biological data arising from frontier technologies. The emergence of these new platforms is generating a vast amount of data with enormous potential to help understand the functioning of the human body in health and disease, as well as the health of animals, plants and our environment more generally. Her expertise in multivariate statistics, combined with her deep understanding of molecular biology, put her at the forefront of cutting-edge biological research. Dr Lê Cao has a track record of success in biological data analysis, in developing novel statistical methods, in implementing them in efficient software, and in disseminating the software and encouraging its uptake by the relevant research community. She plays a critical role in several local, national and international collaborative studies with researchers from diverse bioscience disciplines.

Associate Professor Stephen Leslie, University of Melbourne

Associate Professor Leslie has made major contributions to mathematical genetics. The thrust of his research is developing methods for analysing modern genetic/genomic data, focusing on understanding the role of genetics in human disease and how genetics informs studies of human population history. He applies novel approaches to genetic data to understand the history of populations and infer past migration events. Stephen’s work on the British population is a landmark in the field, impacting history, archaeology, anthropology, and linguistics. It is a blueprint for studies in other populations and a benchmark for understanding natural genetic variation in human populations, crucial for disease studies. In other work he has revolutionised the study of immune-system genes, particularly those crucial to the body’s mechanism for detecting ‘self’ (one’s own tissues) from ‘non-self’ (such as viruses and bacteria), by enabling these genes to be included in large studies for the first time. This work has led to important discoveries associating these genes to serious diseases.

2017

Associate Professor Joshua Ross, University of Adelaide

Associate Professor Ross has made important and influential contributions to Applied Probability and Statistics, and through application to Conservation Biology and Public Health. His research has focused predominately on addressing problems arising in infectious disease epidemiology and conservation biology, though the methodological developments that he has provided to solve such problems are more widely applicable. These application topics are of great importance, and his contributions to these fields are significant. 

2015

Associate Professor Yee Hwa Yang, University of Sydney

Associate Professor Yang is an applied statistician who has made significant contributions to the development of statistical methodology for analyzing molecular data arising in contemporary biomedical research. Her work on removing extraneous variability for microarray data has been incorporated in major software packages used worldwide to identify gene expression patterns. She has also developed novel methods for integrating molecular and clinical data and has already made an impact on Melanoma research by identifying potential genes that help with predicting survival outcome.

2013

Dr Aurore Delaigle, University of Melbourne

Dr Delaigle's has made influential contributions to contemporary statistical problems, including deconvolution, regression with measurement errors, functional data analysis, analysis of high dimensional data, group testing, and a wide variety of contributions to function estimation. She is remarkably adept at transforming complex and highly abstract methods into easy to understand concepts, and at developing fully applicable techniques that work in a wide variety of settings. An advantage of her approach to developing methodology is that her techniques apply at once to many practical problems, in both the biological and physical sciences.

2011

Dr Scott Sisson, University of New South Wales

Scott Sisson has made highly significant contributions to computational statistics and extreme value modelling. His research in approximate Bayesian computation has enabled researchers at the leading-edge of many scientific disciplines to examine realistic models and hypotheses, rather than be forced to use simpler, less credible alternatives. His research on extreme value modelling has enabled improved inferential procedures and highlighted the dangers of poor statistical modelling. In applying these techniques to challenging problems in other disciplines, Scott has had a very positive impact on furthering scientific research in a wide range of applications.

Dr Mark Tanaka, University of New South Wales

Mark Tanaka’s research concerns the evolution and population biology of microorganisms. He uses mathematical and statistical methods to study the dynamics of bacteria and viruses. A particular focus of his research is the transmission patterns of infectious diseases. He has investigated key parameters in the epidemiology of tuberculosis in published research which has led to conclusions with public health policy implications that were hitherto unavailable. Tanaka’s research is highly original and excellent, judged by the highest international standards.

2009

Dr Melanie Bahlo, The Walter and Eliza Hall Institute for Medical Research.

Melanie Bahlo is an applied statistician working in genetics and bioinformatics. She is highly regarded for her work in theoretical population genetics, in genetic epidemiology, and in gene mapping. She began her career with some very strong work in theoretical population genetics. Subsequently she moved into epidemiological applications and has grown into an outstanding statistical geneticist and biometrician.

2007

Professor Robin Hyndman, Monash University, Melbourne

Rob Hyndman has made major contributions to a wide range of fields, especially to forecasting, time-series, graphical methods and computational statistics. His research in forecasting challenged the appropriateness of the most fundamental of Bayesian forecasting models for exponential-family time series and on state-space models for exponential smoothing. Rob has recently proposed a new methodology for forecasting age-specific mortality curves and all official Australian cancer forecasts now use this method. His research on graphical methods has produced an ingenious method for visualising conditional probability densities, and a remarkably useful tool for data analysis – the ‘highest density region’ box plot.

2005—M.W. Blows
2003—N.G. Bean
2001—A. Xia
1997—M.P. Wand
1993—P.K. Pollett
1990—A.H. Welsh