Admission requirements
Assumed prior knowledge
It is strongly recommended that students:
have successfully completed Introduction to Deep Learning.
is interested in scientific research.
is proficient in programming, machine learning, presenting and report writing.
This is an advanced course. We will work intensively on state-of-the-art research topics.
Warnings:.
- Peer feedback is an important aspect of the course, and presence at the weekly meetings is mandatory.
Description
In recent years we have witnessed an explosion of research, development, and applications of Deep Learning. In this seminar we will dive into recent deep learning papers.
The course is structured as follows. The course centers around a group assignment which runs from the very start of the course until the end. In groups you will first select a specific deep learning paper of interest from a top conference in deep learning or related fields, e.g. computer vision, NLP, audio processing, medical imaging. The main goal of the course is to understand the paper, replicate its results and extend/adapt it in some way. Throughout the course you present your progress to the group for feedback and give feedback to others. At the end of the course, you write a report about your results based on the guidlines of a true research paper.
In addition to the specific topic you will dive deep into with your project, you will also gain a wide overview of the current state of this area. This happens through presentations of the other groups and through the lectures. Lecture topics may differ per year as they are selected to be relevance to the project topics and the interest of the students. In 2024-2025 topics included: advances in computer vision, model adaptation, self-supervised learning, data augmentation, model robustness, explainability, multimodal learning, mamba and retrieval augmented generation.
Course objectives
During the course students will:
Understand recent developments in Deep Learning, for instance transformers, foundation models, diffusion models
Analyse scientific papers in deep learning or related fields, e.g. computer vision, natural langauge processing, representation learning. Both understand the content of a paper and identify potential research directions which extend the work.
Run the code associated with an existing deep learning research paper on an HPC GPU enabled research cluster, e.g. ALICE
Adapt the code associated with an existing deep learning research paper implementing some of the potential research directions identified
Evaluate the impact of the new research directions by evaluating their performance on existing datasets.
Understand how to practice effective team work and apply this understanding to the group project
Understand how to give an effective presentation and apply this understanding to the final presentation summarizing the prior work, new research direction implemented and their evaluation
Create a scientific report summarizing the prior work, new research directions implemented and the evaluation of results.
Timetable
The most recent timetable can be found at the Computer Science (MSc) student website.
In MyTimetable, you can find all course and programme schedules, allowing you to create your personal timetable. Activities for which you have enrolled via MyStudyMap will automatically appear in your timetable.
Additionally, you can easily link MyTimetable to a calendar app on your phone, and schedule changes will be automatically updated in your calendar. You can also choose to receive email notifications about schedule changes. You can enable notifications in Settings after logging in.
Questions? Watch the video, read the instructions, or contact the ISSC helpdesk.
Note: Joint Degree students from Leiden/Delft need to combine information from both the Leiden and Delft MyTimetables to see a complete schedule. This video explains how to do it.
Mode of instruction
Presentations, discussions, feedback on students presentations and reports.
Course load
Total hours of study: 168 hrs. (= 6 EC)
Lectures: 10 hrs
Discussions and presentations on student's project: 16 hrs
At home preparation: 142 hrs
Assessment method
The final grade will depend on timely selection and reproduction of the results of a state of the art research paper, quality of presentations and the quality of the final report. Each team will receive a grade which is a rounded average of grades for the following components:
Selection of suitable paper (10%)
Reproduction of paper's results (20%)
Presentation of results (20%)
Final Report (50%)
The team's grade will be internally distributed among team members according to their individual contributions to the project.
To pass the course, the student needs to be present in at least 80% of the lectures. In case of a fail grade (<5.5), there is an opportunity to resumbit the project report, but only with the same research project.
Reading list
The reading list changes every year according to recent developments in deep learning and the project topics chosen by students. See the course page of brightspace for more information.
Registration
Students should also sign up for the course before 15 January 2025. Due to the format of the course, the maximal number of students that can participate in this course is limited to 30. In case when more than 30 students would like to attend this course we will prioritize students on the Artificial Intelligence specialisation.
As a student, you are responsible for enrolling on time through MyStudyMap.
In this short video, you can see step-by-step how to enrol for courses in MyStudyMap.
Extensive information about the operation of MyStudyMap can be found here.
There are two enrolment periods per year:
Enrolment for the fall opens in July
Enrolment for the spring opens in December
See this page for more information about deadlines and enrolling for courses and exams.
Note:
It is mandatory to enrol for all activities of a course that you are going to follow.
Your enrolment is only complete when you submit your course planning in the ‘Ready for enrolment’ tab by clicking ‘Send’.
Not being enrolled for an exam/resit means that you are not allowed to participate in the exam/resit.
Contact
Lecturers: Dr. Hazel Doughty
Website: Brightspace
Remarks
Due to the format of the course, the maximal number of students that can participate in this course is limited to 30.
This course is also suitable/recommended to PhD students who want to use Deep Learning in their projects.
Software
Starting from the 2024/2025 academic year, the Faculty of Science will use the software distribution platform Academic Software. Through this platform, you can access the software needed for specific courses in your studies. For some software, your laptop must meet certain system requirements, which will be specified with the software. It is important to install the software before the start of the course. More information about the laptop requirements can be found on the student website.