Statistics Across Campuses

The Category of Correlations Date: 31 July 2020, Friday Time: 3 pm Speaker:  Prof Annabelle McIver  (Macquarie University)  Abstract: Designing programs that do not leak confidential information continues to be a challenge. Part of the difficulty arises when partial information leaks are inevitable, implying that design interventions can only limit rather than eliminate their impact. We show, by example, how to gain a better understanding of the consequences of information leaks by modelling what adversaries might be able to do with any leaked information. The presentation is based on the theory of Quantitative Information Flow and uses the well-known probability monad to provide an information-flow aware semantics for a small programming language. We will explore some properties of the language and demonstrate that "correlations" rather than the more familiar "prior/posterior" probabilities of Bayesian reasoning are fundamental to understanding how information leak...

Metric number theory via geometry and dynamics: Mahler to Margulis Joint Stochastic and Mathematics colloquium at La Trobe University. Speaker: Dr Mumtaz Hussain, La Trobe University Time & Date: 11:00am Thursday 25 June 2020   Contact the organizer: Andriy Olenko a.olenko@latrobe.edu.au Venue: zoom meeting, see details below Abstract: There are two well-known approaches in solving the measure theoretic problems in Diophantine approximation.  The metrical approach arise from the geometry of numbers and the ergodic theoretic approach arise from the dynamics on the space of lattices. One of the main ingredients in the geometry of numbers is the usage of Borel-Cantelli lemmas from probability theory. Dynamics on the space of lattices rely on the Dani correspondence principle (1985) which was extensively  developed further by Margulis and Kleinbock.  I will discuss both of these approaches and along the way discuss some well-known results such as the resol...

Analysis of repeated categorical ratings: going beyond inter-rater agreement. Statistics and Stochastic colloquium at La Trobe University. Speaker: Dr Damjan Vukcevic, University of Melbourne Time & Date: 12:00pm Thursday 18 June 2020 Venue: zoom meeting, see details below Contact the organizer: Andriy Olenko a.olenko@latrobe.edu.au Abstract: A common task in health and medicine is the classification of patient information into one of several categories by a trained expert. This could include assessing the presence and type of a tumour from a medical image or providing a disease diagnosis from a series of medical tests. Often such judgements are hard to make and error prone: two experts may rate the same scenario differently or the same expert may provide alternative ratings of the same scenario when rating it multiple times on different occasions. Analysing the performance of such expert ‘raters’, and the accuracy of their ‘ratings’ across a series of ‘items’, is a comm...

Estimation of Graphical Models for a class of Multivariate Skew-Symmetric Distributions Date: 26 June 2020, Friday Time: 4pm Speaker: Dr Linh Nghiem (ANU) Abstract: We consider the problem of estimating graphical models for data generated from a class of multivariate skew symmetric distributions, which can be used to model multivariate data with both moderate skewness and heavy tails. Conditional independence between any component requires both the corresponding element of the inverse covariance matrix and the product of the two corresponding components in the shape vector to be zero. Utilizing new properties of the conditional expectation in this class of distributions, we propose a novel two-step nodewise approach to estimate the graphical model. For each nodewise regression, we first fit a linear model using least squares, and then fit a one-component projection pursuit regression on the residual obtained from the first step. The graph is estimated by thresholding an app...

Spatial Confounding in GEEs Date: 19 June 2020, Friday Time: 4pm Speaker: Dr Francis Hui (ANU) Abstract: Generalized Estimating Equations (GEEs) are a popular tool in many scientific disciplines for investigating the effects of covariates on the mean of a response. In the context of spatial analysis, GEEs rely on specifying a regression model for the marginal mean, a variance function, and a working correlation matrix characterizing the spatial correlation between observations. One of the key features of GEEs is that estimation of the covariate effects is robust to misspecification of the (spatial) working correlation matrix. That is, the choice of working correlation only affects efficiency and not the consistency (effectively, the target) of the GEE estimator.  In this talk, we introduce and explore the concept of spatial confounding in GEEs. Specifically, we show that in settings where the covariates included in the GEE are (also) spatially correlated, the choi...

Bayesian modelling of complex trajectories: a case study of covid-19 Date: 12 June 2020, Friday Time: 2 pm Speaker: Prof. Kerrie Mengersen (Queensland University of Technology) Abstract: Since the initial outbreak in Wuhan (Hubei, China) in December 2019, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the virus responsible for coronavirus disease 2019 (COVID-19), has rapidly spread to cause one of the most pressing challenges facing our world today: the COVID-19 pandemic. Within four months of the first reported cases, more than two and a half million cases were confirmed with over two hundred thousand deaths globally, and many countries had taken extreme measures to stop the spread. Although Bayesian models of epidemics are well known in the literature, modelling COVID-19 has been problematic because of the complexity of control responses that were implemented to contain the spread of the disease in different countries. In this presentation, I will describ...

I nferring genetic linkage maps from high-throughput sequencing data Date: 05 June 2020, Friday Time: 2pm Speaker: Dr Matthew Schofield (University of Otago) Abstract: Genetic maps are usually the starting point for many types of genetic analysis. They are one-dimensional representations of genetic inheritance across a chromosome. Genetic maps frequency are commonly inferred from estimates of a hidden Markov model (HMM) since only the expression and not the transmission of genetic information is observed. No general approaches exist for assessing the uncertainty of the map. In this talk, we will obtain genetic maps and associated uncertainty for data arising from high-throughput sequencing (HTS). HTS technology provides high density data from a large numbers of individuals in a cost- and time-efficient manner. However, the observed data from HTS are more error prone than previous technologies. We first extend the HMM to account for error introduced by HTS. We then us...

Social media analysis and COVID-19 Date: 5 June 2020, Friday Time: 3 pm Speaker:  Dr  Lewis Mitchell (University of Adelaide)  Abstract: The COVID-19 pandemic has produced a number of areas where mathematical modelling and data science might make important contributions to the public health response. Concurrently, it has led to a unique improvement in the number of datasets (some anonymised, some not) being provided by typically-ungenerous tech companies to researchers to potentially assist with this response. This talk will explore how we are utilising a few of these datasets coming from the large social media platforms to attack COVID-related problems, including: Measuring social distancing and predicting risk using Facebook data Quantifying the ‘arc’ of patient experience of COVID-19 using Reddit Contact tracing: tracking public sentiment towards the COVIDSafe app using Twitter, and modelling app effectiveness Zoom link:     https://macqu...

The use of Fast Fourier Transforms and Generalized Poissonian distribution to study COVID Deaths Date: 22 May 2020, Friday Time: 3 pm Speaker: A/Prof John Nichols (Texas A&M University) Abstract: The COVID Fatality data is often grouped into subsets that represent political boundaries, if these political boundaries represent unique compact urban areas fully contained in the urban sense then the application of the SEIR model appears to be somewhat applicable, but, if this is not the case, the assumptions that are made for the SEIR model may result in a poor predictive model. The use of Fast Fourier transforms of the residual data from an exponential regression analysis provides a method to estimate the frequency response of the residuals, which can be used to review the SEIR modelling of the urban area.  The second method is a GPD analysis of the daily ratio of the fatalities to the prior day, which may prove to be a method to determine unique compactness. Examples u...

Optimizing the Fitting of Linear Mixed Models - Comparing BLAS Subroutines in Isolation (no pun intended) Date: 22 May 2020, Friday Time: 2 pm Speaker: Luke Mazur (Univeristy of Wollongong) Abstract: Linear mixed models arising from animal and plant breeding result in sparse sets of Mixed Model Equations with particular structures. An effective method of fitting these models is the Average Information (AI) algorithm, and the largest computational bottlenecks in the AI algorithm are the solution of these equations and the calculation of the Sparse Inverse Subset for the AI updating. There are a number of potential Basic Linear Algebra Sublibrary (BLAS) subroutines that can be used for these tasks, and potential candidates are investigated via a comparative experiment to see which combination of these is best. Link: https://uow-au.zoom.us/j/91318598806 Video: