Research project title: Variational Inference for Heteroscedastic and Longitudinal Regression Models
Describe your research project Within the past decade we have been moving into an age where statistics and computing have taken the spotlight. The astronomy and genomics sciences, which first experienced this explosion of information in the 2000s, coined the term “Big Data”.
Currently, the world is dealing with such large volumes of data that it may not fit into standard computer memory and therefore will need to be processed quickly on arrival and then disposed of immediately.
As an example, Facebook, a company that wasn’t around a decade ago, gets more than 10 million new photos uploaded every hour. This helps in creating a digital trail that Facebook can use to learn about it’s users preferences.
A big question that remains however, is:
“How do we attempt to pull the signal from the noise in a timely manner?” An answer that I have been looking into, in my PhD candidature, is Fast Data Analysis.
I have been working on building and implementing fast algorithms that approximate complex mathematical expressions in order to achieve analysis in the quickest time possible. I have seen results in my work being hundreds of times faster than conventional methods, and in the world of big data this would really make a difference.
As speed is achieved however, the quality of the analysis lessens ever so slightly. But, so far my results show strong speed and accuracy combined.
What attracted you to research at UTS Science? The friendly and helpful environment at UTS very much excited me to be a part of UTS Science, along with the quality of research staff that was readily available to help out.
In particular, the research area of my PhD supervisor, Professor Matt Wand, appealed to me greatly and was the main reason for my continuing on to pursue a PhD.
What is your future? Within my PhD candidature I have learnt how important Machine Learning and Computing is to the world of Big Data. As a result, I am now pursuing a post-doctoral training program in Data Science in New York City and hope to eventually apply what I've learnt to a global market in the Data Science industry.