Exam information
Hope you have learnt a lot about Machine Learning throughout your time with us the past two months!
As you are aware, your final exam will happen in Week 11. All the best!
Here is some basic information about your exam:
- The exam will be held in person.
- You will answer all 3 out of 3 questions.
- The duration is 90 minutes.
- You will be writing out your answers by hand in a provided exam booklet.
- You may bring in at most one piece of A4 paper with your own notes (the department’s default). You can write/draw/print as much as you want on both sides, and you cannot bring in a magnifying glass (or equivalent) into the exam.
- Calculators will be provided. You are not allowed to bring your own calculators. I understand that the department will be using a new calculator model this term: FX-85GTCW (subject to successful delivery of goods).
- See the department’s guide for exam candidates for information about exam logistics.
- Past year exams are available on the department’s exam page.
- Papers from 2020/21 onwards follow the latest syllabus and exam structure/style.
- We have two editions of the course (Term 1 and Term 2) since 2019/20. Both share the same syllabus. Both are relevant.
- Ignore the module codes - we are as confused about them as you are! Look for Term 1 (or Term1) and Term 2 instead.
- 2019/20 (C553 and C395) follow the same syllabus, although the exams were “answer 2 out of 3 questions”
- Official model answers are not provided for past year exams (departmental policy). Instead, official examiner feedback is provided since Term 2 2019/20:
- Examiner feedback for 2019/20 (term 2 only)
- Examiner feedback for 2020/21 (both terms)
- Examiner feedback for 2021/22 (both terms)
- Examiner feedback for 2022/23 (both terms)
- Examiner feedback for 2023/24 (both terms)
- Be cautious about the student-contributed answers in the (un?)official “Secret Revision Folder” to which you probably have been given access. We have personally seen mistakes in the answers in the folder!
- Exam preparation tips:
- Practise past year papers (most important)
- Do tutorial exercise worksheets
- Review courseworks (especially coursework 2) - make sure you understand all bits of your code (whether you were the one who implemented them)
- The Google Colab lab tutorials may help deepen your understanding
- The quizzes may also help deepen your understanding of the topics
- FAQ: How many decimal places are needed for the calculations?
- At least 3 significant figures recommended (more is fine)
- Please provide workings for any calculations
- Especially important if your answer varies too far from the sample answer. The examiners will examine your workings to check whether they are just rounding errors (which is fine), or whether there was actually a mistake in the calculations (in which you will be penalised).
- The instructors will stop answering questions on the Ed board 24 hours before the exam.