Meet the Speaker – Leah South

We talk to Dr Leah South from Queensland University of Technology about learning new things and the next big questions she hopes to tackle in her field of research.

1. What drives your interest in this field?

What drives me is anticipation for the rare moments when you develop a new method and it actually works! I’m particularly interested in Bayesian computational statistics because it’s a rapidly developing area so I learn something new most days.

2. What are the next interesting “big questions” or challenges you hope to tackle?

So-called “biased” MCMC methods for big data are promising in that they offer the potential to get reasonable estimates of quantities of interest in very little time. Unfortunately, practitioners are wary of using these methods when they don’t know if they can trust the results. There has been a lot of interest recently in assessing quality in this context but the existing methods can be very slow, meaning that you lose some of speed benefits you had from using biased MCMC in the first place. At the moment, I’m trying to develop a fast alternative to these methods.

3. If not mathematics research, what would have been?

I was thinking about going into medicine in grades 8-10 but I was too put off by the need for memorisation! I preferred maths where a thorough understanding often means that you can recall the concepts from first principles rather than rote memory.

4. Deepening field knowledge and providing a networking platform, why are opportunities such as AMSI Winter School so valuable? What do you hope attendees take from your lectures?

AMSI schools offer fantastic opportunities to learn about cutting edge methods and to network. I met a collaborator on one of my most exciting research papers through a similar event.

I’m hoping that attendees can take away a toolbox for assessing and improving the performance of their MCMC algorithms through efficient post-processing.

Dr Leah South will be presenting the topic Post Processing of MCMC at the 2021 Winter School at Queensland University of Technology from 12-23 July