TWiki> ANC Web>LfdCourseChangeComments (05 Dec 2007, Main.s0451403)EditAttach

Changes to Machine Learning Courses

We are currently contemplating significant changes to the machine learning program. One potential proposal (being contemplated, but by no means decided) is to split LFD into two courses. The first course (Sem 1) would teach a number of machine learning techniques, focus on using the tools but without going into as much detail about why these tools work. There would be a coverage of a wider range of tools than currently given on LFD, and more practical examples of the tools in action, with examples taken from a number of domains including language modelling. However it would lack the detail needed for a good general and rigorous understanding of machine learning. This is likely to be a level 9 course possibly only available to MSc students, as it will overlap significantly with the undergraduate Inf2B syllabus.

The second course (Sem 2) would be a more technical course, somewhat along the lines of LFD, covering the reasons why machine learning methods work, the methods in detail, maximum likelihood, and updating things to spend more time on things beyond maximum likelihood, such as Bayesian methods. Hence it would be more cutting edge than the current LFD course. There would still be a practical element to the course. Possibly the overlap with PMR could be reduced as PMR would already have been taught. The ideawould be that the first course would be most suitable to those minoring in machine learning, and the second for those majoring in machine learning. You would potentially be allowed to take both courses. DME would have PMR as a prerequisite and the second machine learning course as a co-requisite.

We would welcome your opinions on this. Specifically

  • Would it have served you better if you had been presented with these options instead of the current ones or would it be worse? Why?

  • What would you have chosen to do if these were the options available?

  • If you would want to take the second semester course, would you also have taken PMR (ie if we made PMR a prerequisite would it have been a problem to you)?

You can make your comments on these issues or anything else that occcurs to you about this matter here. (Please use a mozilla browser if accessing from outwith DICE). Thanks for your comments. They will be a great help to us. Comments need to be made by Friday 16th Nov, as the review will happen very shortly. The more comments we get the better a feel we will get for the demand. So please say something even if you don't care and don't feel strongly in any way.

Your Comments

Would it have served you better if you had been presented with these options instead of the current ones or would it be worse? Why?

  • My specialism is more on the bioinformatics side of Informatics. I like learning both application and theory, but I don't think I would want to use up more than one of my module choices on machine learning. If the application/theory were to be split, I'd end up missing out on one or the other.
  • I'm doing a joint degree in Maths and AI. I'm taking four informatics courses this year (4th year) and so to use two of my choices for LfD would have been prohibitive. I would have chosen something else instead. If the first new course was not a prerequisite for the second, I would have just done the second. I'm glad I was able to do LfD this year, because now I can leave university with a proper understanding of how machine learning works.

What would you have chosen to do if these were the options available?

If you would want to take the second semester course, would you also have taken PMR (ie if we made PMR a prerequisite would it have been a problem to you)?

Your General Comments...

Topic revision: r4 - 05 Dec 2007 - 17:12:06 - Main.s0451403
 
This site is powered by the TWiki collaboration platformCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback
This Wiki uses Cookies