Statistics Across Campuses

Wasserstein Regression for Distributions and Distributional Time Series Date: 2 December 2021, Thursday Time: 10-11am AEDT  Speaker:  Professor Hans-Georg Müller  ( University of California, Davis ) Abstract:  The analysis of samples of random objects that do not lie in a vector space has found increasing attention in statistics in recent years. An important setting for distributional data analysis are samples consisting of univariate probability measures defined on the real line. Adopting the Wasserstein-2  metric, we propose a class of regression models for such data, where random distributions serve as predictors and the responses are either also distributions or scalars. To define this regression model, we utilize the geometry of tangent bundles of the metric space of random measures with the Wasserstein metric. The proposed distribution-to-distribution regression model provides an extension of classical linear regression to the case of distributional data. We study asymptotic rate

  The Moyal Medal Committee invites you to attend the virtual lecture and presentation of the 2021 Moyal Medal to  Professor Louise Ryan  from the   University of Technology Sydney   The Moyal Medal is awarded annually for research contributions to mathematics, physics or statistics, the areas of research of the late Professor José Enrique Moyal. Professor Moyal was Professor of Mathematics at Macquarie University for five years from 1973 to 1977. His insight into the interaction between mathematics, physics and statistics led him to make contributions to these disciplines which have had far-reaching ramifications in all three fields.  2021 Medallist Professor Louise Ryan  After completing her honours degree in statistics at Macquarie University in 1978, Louise Ryan left Australia to undertake PhD studies at Harvard. On completing her PhD in 1983, she moved to the Harvard Biostatistics Department, first as a postdoctoral fellow, then a junior faculty member and finally as the Henry Pic

BERT-based language models for US Elections, COVID-19, and analysis of the translations of the Bhagavad Gita Date: 26 November 2021, Friday Time: 4pm AEDT Speaker: Dr Rohitash Chandra (UNSW Sydney) Abstract: There has been tremendous progress in language modelling with deep learning via long short-term memory (LSTM) models and variants such as bidirectional encoder representations from Transformers (BERT). Motivated by these innovations, in this seminar, we present three multi-disciplinary applications of BERT-based language models. Firstly, we develop a framework to model the US general elections. We investigate if sentiment analysis can provide a means to predict election outcomes. We use BERT-based language models for Twitter sentiment analysis leading to the US 2020 presidential elections. Secondly, we present a framework for sentiment analysis during the rise of novel COVID-19 cases in India. Finally, we compare selected translations (mostly from Sanskrit to English) of t

           PCA, likelihood ratios, and a transition to the Tracy-Widom law Date :  Thursday, 18 November 2021 Time:   10 am  - 11 am Speaker:  Professor  Iain Johnstone  ( Stanford University and ANU ) Abstract: The Tracy-Widom distribution has found broad use in statistical theory  and application. We review some of this, focusing first on Principal Components Analysis and the `spiked model'. When the spike signal is below the Baik-Ben Arous-Peche threshold, likelihood ratio tests for presence of a spike are more efficient than the largest eigenvalue in many settings of multivariate statistics. In recent work with Egor Klochkov, Alexei Onatski and Damian Pavlyshyn, we study the likelihood ratio test in the transition zone around the BBP threshold, making use  of a connection with the spherical Sherrington-Kirkpatrick model. Zoom Link:  https://anu.zoom.us/j/86015752944?pwd=TEpvRjRrTDh0NWZMc0J3RDVNVndHQT09 Meeting ID: 860 1575 2944 Password: 069205 Please contact Yanrong Yang ( yan

          High-dimensional MANOVA via Bootstrapping Max Statistics Date :  Thursday, 11 November 2021 Time:   11 am  - 12 noon Speaker:  Professor Zhenhua Lin  (National University of Singapore) Abstract: In the talk I will present a new approach to the problem of high-dimensional multivariate ANOVA via bootstrapping max statistics that involve the differences of sample mean vectors. The proposed method proceeds via the construction of simultaneous confidence regions for the differences of population mean vectors. It is suited to simultaneously test the equality of several pairs of mean vectors of potentially more than two populations. By exploiting the variance decay property that is a natural feature in relevant applications, it is possible to achieve dimension-free and nearly-parametric convergence rates for Gaussian approximation, bootstrap approximation, and the size of the test. The proposed methodology, demonstrated with ANOVA problems for functional data and sparse count data,

Biostatistics and SARS-CoV-2: research, policymaking, and communication. Date: 25 November 2021, Thursday Time: 4pm AEDT Speaker: Prof Geert Molenberghs (KU Leuven) Abstract:  The COVID-19 pandemic, induced by the SARS-CoV-2 virus, is literally a rare event in the course of history. We need to go back to 1918 for an even worse pandemic, the Spanish Flu, or H1N1, although we also had the tuberculosis pandemic in the interbellum; there was the Russian flu in 1890 (maybe also a coronavirus and not influenza), and the plague that literally haunted the world for several centuries. When there are no antiviral means to speak of, and in the absence of vaccines, time-honored non-pharmaceutical interventions enter stage. Apart from controlling the epidemic, for better or for worse, they generate side effects, for society, its well-being, and for the economy. Based on data and imperfect evidence, the biostatistician contributes to understanding what has happened and is happening, is able to separ

Astrostatistics in Gravitational-wave Astronomy Date: Tuesday, 9 November 2021   Time: 7:00pm (AEDT)     Speaker:  Ilya Mandel (Monash University, Australia)  Contact the organizer: Andriy Olenko a.olenko@latrobe.edu.au Abstract: Modern astronomical data sets often raise challenges associated with selection biases, accounting for confusion between backgrounds and foregrounds, and performing inference on big data with complex, multi-parameter models. I will discuss some of the techniques that we used to attack these problems, illustrating them with results from gravitational-wave observations of merging black holes … and a bit further afield. See zoom meeting details below. This seminar is a part of a new international online seminars series on Statistics and Data Science applications in Astronomy : https://sites.google.com/view/iau-iaa-seminar/home This seminar is an initiative of the International Astrostatistics Association and the IAU Astroinformatics and Astrostatistics Co