Skip to content

Syllabus

Last updated: 21 September 2026

Schedule

The schedule may be adjusted slightly to match the pace of the class. All sessions, including the exam slot on 21 December, meet at the normal class time and in the normal room. Slides, notebooks, readings, and submission links will be added as they become available.

Date Session Topic Slides Lab Homework
7 Sep S1 Orientation and foundations Slides Survey Prerequisites
14 Sep S2 AI-assisted research workflow Slides Notebook · Sol n/a
21 Sep S3 Understand your data Slides Notebook n/a
28 Sep — No class — Teachers' Day n/a n/a Notebook
5 Oct S4 Regression, loss, and optimisation
12 Oct S5 Classification: kernels, trees, and boosting
19 Oct S6 Neural networks
26 Oct — No class — Retrocession Day n/a n/a
2 Nov S7 CNNs and inductive bias
9 Nov S8 Sequence models and time series
16 Nov S9 Consolidation and invited seminar
23 Nov S10 Attention, transformers, and sets
30 Nov S11 Autoencoders and variational autoencoders
7 Dec S12 GANs, diffusion, and invited seminar
14 Dec S13 Final presentations I n/a n/a n/a
21 Dec Exam slot Final presentations II n/a n/a n/a

Assessment

Component Weight
In-class labs and weekly homework 60%
Final project and presentation 40%

In-class labs and homework are set as 10–12 assignments across the semester:

  • due at noon on Saturday; and
  • late submissions are not accepted.

In-class labs count as attendance, so taking part is what is marked — you are not expected to finish every part in the session. Get as far as you can with your group and complete the rest in your own time. Missing a session means no score for that lab, unless you have a good reason such as illness or a family emergency: tell the instructor or a teaching assistant and it will not count against you.

Attendance at the final presentations

The final project presentations on 14 and 21 December are compulsory. Absence without prior permission costs 10 points — 10% of your final grade. If you cannot attend either date, tell the instructor in advance.

Bonus

An optional competition will run during the semester, offering extra credit on top of the marks above. Details coming soon.

Use of AI-assisted coding

AI-assisted coding is strongly encouraged, but all generated work must be verified. You remain responsible for the correctness of your code, analysis, citations, and conclusions.

AI use must be disclosed in the submitted results.

What a disclosure looks like

A sentence is enough. Say which tool you used and what you used it for:

This website was built with the assistance of Claude Code, for drafting, formatting, and grammar checks.

The training loop was written with GitHub Copilot; I rewrote the loss function myself after the generated version failed a dimensional check.

The first is not hypothetical — it is the disclosure for this course's own site and slides, on the home page.

Tools and accounts

  • Accounts: complete all required accounts by Session 3. Course services, notebook-computing options, and NSTC Core limits are on the Resources page.
  • Software: installation is covered under Software setup.
  • Version control: Git is used throughout the course.

Collaboration and submissions

  • In-class notebook labs are completed in assigned groups of three students. Groups rotate every week; check the teaching assistants' announcement for that week's assignment.
  • Submit one set of work per group: a single in-class lab and homework submission covers all three members. Put every member's name on it, with a short note on what each person contributed.
  • Accommodations and exceptional circumstances should be discussed with the instructor as early as possible.