Welcome machine learning students!
ML-Base is a time and space for machine learning (ML) students taking IAML, MLP, or MLPR to meet, work together, and sometimes get advice from someone who is doing ML research.
Appleton Tower 5.04 is your space every day 5–7pm, Monday to Friday (weeks 1–11 Sem 1).
If you need a table, you can ask any non-ML students to leave.
Contents: Start of Semester — How to use ML-Base — Schedule
There is currently no sign-up system, as a manageable number seem to be coming naturally each day. If that changes, we may have to do something to limit numbers. Please don't wait until close to an assignment deadline to come!
In week 1, an ML lecturer will attend ML-Base 5–6pm on Tue/Wed/Thu (17–19 Sep) to give advice on course selections. Please read the advice in the MSc handbook and the MLPR background notes first. Tutors will discuss your tutorials and coursework from week 2.
If you already know a lot of machine learning, please don't swamp lecturers with advanced questions in week 1, while other students are trying to finalize their course choices.
ML-Base could be crowded at first. If so, meet some other students there, and find somewhere else in Appleton Tower to sit. If you want to talk to a member of staff, you may have to queue or try the next day. Over time, we expect numbers to even out over the week. If not, we'll implement a sign-up system if we have to. Please be patient while we get the balance right.
Motivation: Employers primarily hire graduates because they are looking for people who ask the right questions, can explain their reasoning to other people, and get on top of new technical tools quickly. They may also be looking for particular specialist knowledge, but it is unlikely whether they will care that you can memorize the answer that someone told you for question X of tutorial Y!
ML-Base is meant as a friendly space for you to work out what you need to know for your courses. But it won't be useful to just be told answers. In the future (and your exams and assignments) you will need to be able to work out and explain new things. Therefore you should get as far as you can through active work, discussing the course materials with those on your course.
Some of you will find that you're ahead of other students in some areas. Some of you have done more mathematics, others more programming, others more actual machine learning. We hope you'll still find ML-Base useful: explaining something to others is a valuable skill, and probes if you really know it. If you are in the privileged position of being ahead in some ways, please be mindful of the space you occupy in conversations, and how you might make other people feel. We want ML-Base to be a friendly, mutually-respectful place, where everyone gets their turn to speak.
Where your groups are still stuck, the tutors may be able to help; they are researchers in machine learning and so have broad expertise. However, they often won't have worked through the precise questions that you're working on. They can still help you, but you will need to explain to them what you're doing, and why you are stuck. Often they won't just tell you the next step — if it's a hard question, they might not even immediately know it — but they will be able to advise you on how to think about a problem, and what you might do next. If you can explain where your group is finding a course difficult, they can also feed that back to lecturers, which is really useful.
Hopefully it's now clear that ML-Base is not a lecture or a tutorial, and tutors aren't going to give you the answers to assignments. Don't come and expect to sit back and listen. Only come if you're willing to work actively with other people. Tutors will need to circulate, so can't sit with you to work on all your questions. Try to use their time effectively by working out everything you can with others on your course.
We have five ML researchers who will attend ML-Base and help groups of students where they can. They will start by attending on the days shown below from week 2 (23 September), probably until the end of November (to be confirmed):
| Monday: | Nick Hoernle |
| Tuesday: | Matt Rounds |
| Wednesday: | Maximiliana Behnke |
| Thursday: | Aidan Marnane |
| Friday: | Samuel Haynes |
This schedule is likely to change from week to week according to demand and availability. Use ML-Base or lose it! Don't just come before an assignment is due, or at the end of the course. Your studies and ML-Base are about steady engagement during the course, and you're unlikely to get timely help if you come at crunch points.