Exams will be announced on InfoStud.
Booking is mandatory (and it is normally allowed until a week before the exam).
The exam score would be composed of two parts, counting each 50% of the final mark:
An exam on the theory (written or oral, TBA)
A practical part, which consists of
Assignments in Python/Pytorch, communicated during the course and to be submitted by the given deadlines during the course, counting for 2/3
A final project and its presentation, counting for 1/3
Important notes:
The assignments and the final projects must be submitted in groups of size [2-4]. The first lectures are an opportunity to team up in person and/or via discussions on the Google group
In order to take part in 1, it's necessary to achieve a pass on part 2 of the exam
If you pass part 2, you may book the exam part 1 during the course of the next calendar year, i.e. scores in part 2 would remain valid for a calendar year
Bonus points may be given during part 1 at the discretion of the lecturer.
Algorithms, objectives and topics for the final project may be freely chosen within the course material. Project topics would be discussed in class, where sample/potential topics would be illustrated.
Plagiarism is severely prohibited and, in any form, is regarded as a serious violation of the ethical code of conduct. Plagiarism includes the submission of an assignment or project whose source code or report bears strong resemblance to another persons's source code or report, including other AML projects and/or resources that can be found online. After submission, every project would be checked against plagiarism, including automatic detection tools. Assignments and projects resulting incurring in plagiarism would be invalidated.