My MLSS 2009 Lecture

In late January I gave a lecture at the 2009 Machine Learning Summer School in Canberra. A recording of all three hours(!) of it is now available online thanks to videolectures.net:

The material in the lecture was taken from the recent work Bob and I have been doing on information, divergence and risk. As the summer school is aimed at new PhD students, I tried to pitch the material at a suitable level. Upon reflection, I think I probably aimed the level of the material a little too high. For example, I could have spent more time explaining what a Bayes risk was and why it is so important. I also feel as though I rushed over some of the later material, partly because I realised that I had incorrectly assumed the audience’s background knowledge and partly because I was running out of steam at the end of three hours.

It seems that the slides for my talk are not currently available from the videolectures.net site. For those interested, I’ve made a PDF copy available here to make it easier to follow my discussion in the video:

There is a fairly comprehensive list of references at the end with links to the key papers that I talk about.

Mark Reid April 4, 2009 Canberra, Australia
Subscribe: Atom Feed