This person does not currently hold a position at QUT.
Until recently I was a PhD student at the Australian National University and NICTA, studying machine learning under the direction of Prof Bob Williamson. I am now working with Prof Peter Bartlett at both QUT and UC Berkeley on the theoretical underpinnings of machine learning, developing new scalable algorithms for solving machine learning problems as well as proving theorems regarding their performance. In particular, I am interested in the problem of learning from data of variable quality and form.
In the usual theoretical analysis of machine learning algorithms, it is assumed that the data used to train our model of the real world, is of the same quality and form as the data we use to test our model’s conclusions. In practice this is rarely the case. Usually, the “training” data is of lower quality than the “test” data. In my PhD I sought to understand this problem. I devised means to learn from and place value on data of varying quality and form. Below are two recent publications on the topic, that show how my ideas can lead to simple, robust means of learning classifiers.
If this sounds like something you are interested in, I am your guy. Pointers to my work can be found at,