Mini-project report

The mini-project is carried out in groups of 3 to 4 students.

Project overview

The project is to estimate the skill of players, involved in a competition, based on the results of matches between pairs of players. The project follows the whole process of solving a real-world problem using probabilistic machine learning:

1. define a model of the phenomenon of interest,
2. analyze the model to unveil structure,
3. formulate an algorithm to make inferences,
4. use the inferences to create predictions.

Using the techniques taught in the course, you will implement state-of-the-art machine-learning methods based on the Trueskill™ ranking system developed at Microsoft for online matchmaking. You will then apply the methods to rank the teams in the Italian Serie A elite football (soccer) division.

Instructions

All instructions are available in the file APML2023-project.pdf (updated 12/09/2023) Download APML2023-project.pdf (updated 12/09/2023)

Dataset

You can download the dataset of results of the Serie A division 2018/2019 here Download here.

Important dates

The deadlines and other important dates for the project are as follows:

Moment	Deadline	For whom	Comments
Group registration	September 4	Students
Verify model via quiz	September 13	Students	Not mandatory, only recommended deadline to be in "phase" with the project.
First report submission	October 6 @ noon 11:59 AM*	Students
Peer-review submission	October 11 @ noon 11:59 AM*	Students
Second report submission	October 20 @ noon 11:59 AM*	Students	Both revised report based on peer-reviews, and contribution statement
Feedback and grade (pass/revise) from teachers	November 3 @ noon 11:59 AM*	Teachers	Both report and peer-review report
Revised peer-review submission	November 10 @ noon 11:59 AM*	Students	Only if revision on peer-review is required
Final revised report submission	November 17 @ noon 11:59 AM*	Students	Only if revision on mini-project report is required
Feedback on revised report	TBA	Teachers

* Note, all submissions should be done at the latest at noon 11:59 AM the corresponding day.

Checklist before first report submission

Please ensure that:

• Make sure that your project group as registered in Stuidum is consistent with those who have contributed to the project. If not contact us well before (several days) the deadline.
• You have read the final version of the report from start to end, and made sure it is readable.
• The report does not contain material copied from elsewhere (all reports are checked for plagiarism using Ouriginal).
• The report is anonymous (that also goes for the file name).
• The report is written using the NeurIPS style, and is not more than 8 pages long (not including any reference list or possible appendix).
• The pdf of the report is not placed inside the zip-file containing the code.

Submit the pdf and the zip via the "start assignment" button on the top right of this page (if you don't see the button you are most likely not logged in with your Studium account). Submit the two files in one submission (by adding the second file with "add another file").

Checklist before second report submission

All items for the first submission apply, except that...

• In contrast to the first submission, the submission should not be anonymous. That means you should add your names to the author list in the document.

In addition to the items in the first submission

• You have written a few sentences about how you have updated your report based on the peer-review. These sentences are submitted in the comment field when you submit your updated mini-project report.
• You have written the contribution statement in a separate document which is to be submitted here.

Questions

If you have questions, write in the discussion forum or ask your question on the helpdesks sessions.