<aside> ❓
Deadline is 2026-01-30@18:00
Late submission? 10% will be removed for each day that you are late (3 days max)
It’s worth 65% of the final mark
This is a group project (3-4 per group).
</aside>
Before Thursday 18 December at 17:00, please provide a 1-page project plan that summarises your project and presents the main steps and deliverables.
You can use the MoSCoW method. We know you cannot plan precisely, but we expect you to have examined the data and have an idea of what you will deliver.
On Friday 19 December during the session, your assigned supervisor will briefly discuss this with you.
Submit this 1-page plan by email to all three supervisors: Hugo (<[email protected]>), Gina (<[email protected]>), and Hidemichi (<[email protected]>).
| Dataset(s) | 25% |
|---|---|
| Code on GitHub | 25% |
| An engineering report | 25% |
| A final presentation | 25% |
The produced dataset(s) hosted somewhere and accessible. We can arrange for you to have a https://source.coop account.
Code on GitHub publicly accessible (make sure it’s a good repo), and result/demo is hosted somewhere publicly accessible (to be arranged).
A final presentation 15min where you explain what you’ve done and how you solve the issues, followed by 10min questions from everyone.
For the final engineering report, we expect a scientific and well-structured document that clearly defines the problem or research question addressed, explains the methodology and solutions you developed, and critically evaluates their strengths and weaknesses. Your report should include a thorough discussion of alternative approaches or potential improvements, supported by evidence or references where appropriate. Aim for clarity, depth of analysis, and a coherent argument that demonstrates independent thinking and understanding of the topic. Write it for the other teams, so that they could re-implement what you did by reading your report (and having access to your code). We expect 15-20 pages max.
<aside> <img src="/icons/warning_orange.svg" alt="/icons/warning_orange.svg" width="40px" />
If the project involves benchmarking then the dataset(s) part is perhaps less relevant, and we will adjust the weights. We’re flexible.
</aside>