Machine Learning and Pattern Recognition (MLPR)

Response to 2016/17 survey feedback for MLPR.

I have thought a lot about the feedback from students for this course. You can read some of the feedback in the survey results that are published online. However, I've also had feedback from a mid-semester survey, and many conversations with students over the last couple of years.

I completely over-hauled MLPR last year (2016/17) in response to student feedback, and historically high failure rates. Over-all the response was really positive: student satisfaction and marks went up a lot, and there was a real improvement in the quality of exam answers. I'll continue to work on the course and its materials, but I don't expect to make dramatic changes this year.

You will find comments both from students who found the course too hard and others who found it too easy. I will continue to do what I can to include extra material for keen students. I will also continue to try to improve the materials. However, there is a limit to what I can and should do. Those who can't follow the materials or keep up with the course, should take IAML instead.

Did I mention that you should look seriously at whether you should take IAML instead of MLPR? There is a lot more discussion and information on this point in the MLPR class notes, which are openly available online.

For those with the background, I really look forward to teaching this course again, and meeting some of you soon.

Iain Murray, September 2017