My current research interests include: loss functions in statistical and online learning, and connections between learning theory and market mechanisms.
My research vision can be distilled down to wanting to build theoretical and practical foundations for wide-spread, machine-assisted inference. I am doing so through the following funded projects:
July 21, 2015 – Stephen Gould and I gave the first lecture for our new course, The Craft of Computing. Despite a couple of technical glitches I think it went pretty well and it looks like we have a keen cohort of students too.
July 16, 2015 – I gave a seminar to NICTA’s Machine Learning Group. I presented summaries of the work I presented at COLT (with Rafael Frongillo, Bob Willimson, and Nishant Mehta) on generalized mixability and some more recent work with Raf on risk networks.
July 9, 2015 – Tim van Erven, Peter Grünwald, Nishant Mehta, Bob Willimson, and I have just finished a long paper that relates various conditions for fast rates in online and statstical learning. The pre-print is available on arXiv.Older News →
I am developing and lecturing a new first year course, The Craft of Computing (COMP1040) with Stephen Gould in semester 2, 2015. I am also convening a course on Information Theory (COMP2610) which will be delivered by Aditya Menon and Xinhua Zhang in semester 2, 2015.
I previously taught COMP2610 with Edwin Bonilla in 2011 and 2012, and with Aditya in 2013 and 2014. I also co-lectured Advanced Topics in Statistical Machine Learning (COMP8650) in semester 2, 2012. Some course material is available here:
I am currently primary supervisor for three PhD students: Avi Ruderman, Nicolás Della Penna, and Finn Lattimore. I am also supervising two honours students: Andrew Wrigley (RSCS) and Zac Cranko (co-supervising with Simon Grant at RSE).
I am very pleased to announce that Mindika Premachandra submitted her PhD in September 2014 which has been reviewed and accepted with minor corrections.
I am currently the Production Editor for the JMLR Workshop & Conference Proceedings. I am also an Action Editor for the JMLR's Machine Learning and Open Source Software track and the Research Chair for the 2015 conference on Predictive APIs.
I was local organiser for ALT 2010 as well as a member of its Program Committee. I have also been on program committees for ICML, NIPS, UAI, ALT, ACML, and various NIPS and ICML workshops. I have reviewed for journals (JMLR, JAIR, PAMI, TCS, Trans. on Info. Theory, Trans. on NNLS, PloS ONE, AMS Mathematical Reviews, Bernoulli, and the IEEE Procs. on Big Data) and conferences (including NIPS, ICML, COLT, UAI, and AAAI).I have also participated in the Mathematicians in Schools program as a partner with Yass High School.