Week 6, 26–30 October

General warning: the UK’s timezone changed on Sunday 25 October. We are now working on GMT (or UTC), whereas we were previously on BST=UTC+1. If you are not in the UK, you will need to make sure you know the new time difference.

Welcome to week 6! This week we complete Gaussian processes, then reflect on fitting non-Gaussian models.

This is the fourth week to be assessed. Like before, your mark for this week depends on completing the discussion task (10%), attempting the in-note questions (20%), and this week’s assessed questions (70%). Full details on the assessments page including rules you must know.

Reminder: there will be no new material or tutorial meetings for MLPR in week 7. Use the time for review, and to prepare for the class test on Monday 9 November.

Office hours:

You can meet us on MS Teams in the Meet-up channel of the MLPR 2020/21 Chat team on Friday at 9:30 AM and 4:30 PM UK time (GMT, UTC+0 from this week!). One of Arno or Iain will be there. If you want to discuss something individually, please contact us by email: Arno aonken@inf.ed.ac.uk or Iain i.murray@ed.ac.uk.

Distinguished lecture:

11am, Friday 30 October Ross Anderson - The Sustainability of Safety, Security and Privacy

Here is what you need to do in Week 6:

• Any catch-up: The notes and a video this week encourage you to do some review. You might choose to do some of that review in Week 7, ahead of the class test. If there are any threads on hypothesis that didn’t get resolved (allow 48 hrs), email Arno and Iain.

• Lecture notes: Work through the Week 6 notes, answering all the questions. You should answer these as you go, to get the instant feedback and discussion of the answer. It’s fine to make mistakes here, but an honest attempt at these by Friday at 4pm (UK time) is required.

• Question sheet: Do the week 6 question sheet. This question sheet is assessed and forms the bulk of this week’s marks.

• Tutorial group discussions: Either a) Provide a prioritized list of things you’d like to ask your group and tutor about; or b) if your group has no problems, you can discuss the following (challenging) task:

  If we want to express uncertainty about predictions we could do it in non-Bayesian ways. For example, we could fit a second regression model to predict the square residual (or its log). The square residuals of our first model are new labels for the second regression task. What are the advantages for each of: i) this method; ii) Bayesian linear regression; and iii) GPs? Ideally, can you think of situations where you would prefer each of them?

  Your tutor will need to see you either actively discuss difficulties you or others in your group have with the material, or make a reasonable attempt at the task. You can attempt both if you deal with difficulties quickly. Create a short summary of your conclusions for your tutor. This summary is (lightly) assessed! See the group instructions for details on how to submit the group discussion report.

We recommend that you aim to finish the questions (in the notes and question sheet) and submit your discussion report by the end of Thursday. We will assess only what you have submitted by 4pm UK time on Friday.

As in the Informatics late policy, extensions are not available for weekly hand-ins. We expect many students to miss or under-perform on one hand-in, and will discount the one with the lowest mark. If you experience more significant disruption to your studies, you may need to file special circumstances. Consult your Personal Tutor or Student Support team. Lecturers on a course cannot make allowances outside these procedures.