Statistics Across Campuses

A gentle introduction to string-count distributions in random texts  Date: Thursday 25 February 2021 Time: 10am Speaker: Dr Ben O'Neill (ANU) Abstract: An interesting problem that arises in genetic analysis and other contexts is determining the exact or approximate probability distribution of the “string-count” giving the number of occurrences of a fixed character string in a random “text” (i.e., a random vector of symbols). The simplest case of interest is when the symbols in the text are IID categorical random variables. The prima facie simplicity of this problem sometimes attracts the attention of novice analysts, buoyed by the fact that it is easy to compute the probability of a match in the simple case where the length of the text is the same as the length of the string. (And of course, if the text is shorter than the string then the problem is trivial!) However, when the length of the text is larger than the length of the string, the problem becomes complicated, owing to ...

My experience of Bayesian Inference and Deep Learning Date: 26 February 2021, Friday Time: 4pm AEDT Speaker: A/Prof. Richard Xu (University of Technology Sydney) Abstract:  In a research environment where deep learning methods dominate, Bayesian framework still plays a vital role. In this talk, I will first illustrate Bayesian inference through a small tutorial. Then I will describe some of the latest research results (including some of my own works), in which Bayesian inference has been used to explain and assist (and assisted by) deep learning. Finally, I will give some thoughts on the future of Bayesian inference. Bio: Richard Yi Da Xu is an Associate Professor at the University of Technology, Sydney (UTS). He leads a team of 25 people, includes postdoc, PhD students, and data engineers; His primary research is in machine learning and artificial intelligence. Richard published at many top international conferences, including AAAI, IJCAI, ECAI, ECCV, AI-STATS and ICDM as well as...