Login for PhD students/staff at UCPH      Login for others
DTU, DIKU & AAU Summer School on Geometric Deep Learning
Provider: Faculty of Science

Activity no.: 5185-21-02-32 
Enrollment deadline: 17/08/2021
Date and time16.08.2021, at: 08:00 - 20.08.2021, at: 16:00
Regular seats10
Activity Prices:
  - Participant4,700.00 kr.
  - Deltager/Participant Others1,000.00 kr.
LecturersJon Sporring
Stefan Horst Sommer
Raghavendra Selvan
ECTS credits3.00
Contact personJon Sporring    E-mail address: sporring@di.ku.dk
Enrolment Handling/Course OrganiserJon Sporring    E-mail address: sporring@di.ku.dk
Written languageEnglish
Teaching languageEnglish
Exam requirementsECTS points are obtained by presenting a poster at the start of the summer school as well as participating in the group examination on the final day.
Exam formAndet/Other
Criteria for exam assessmentPosters should also be emailed to Patrick Møller Jensen, patmjen@dtu.dk, along with the poster pitch slides, no later than August 6 2021.

Deep learning and especially convolutional neural networks have achieved unprecedented results in a wide range of application areas ranging from image segmentation to language understanding. Most of these results have been obtained in Euclidean domains such as signals and images sampled on a regular grid, however many problems are more naturally modelled in non-Euclidean domains such as graphs and curved manifolds, where convolution is not easily performed. Geometric Deep Learning is a young field concerned with the problems of Deep Learning techniques to graphs and manifolds, and this course will give an introduction to geometric deep learning and touch on some of the exciting new research presently being performed by the topmost researchers in the world.

This course will feature lectures before lunch with exercises to be carried out after lunch. The students will be expected to read a set of given research articles on geometric deep learning prior to attending the course. The students are expected to write a report or submit a poster which details their research with a view towards applying geometric deep learning to their own research.

Learning outcome
- Knowledge about state-of-the-art geometric deep learning theory
- Experience with programming geometric deep learning algorithms

Workload total 70 hours

Course Preparation 15,00
Lectures 34,00
Theory exercises 20,00
Poster presentation 1,00

PRIMARY SIGN- UP AND PAYMENT must go via this site Deadline 6 AUGUST !!

DIKU PhD Students must sign up here also:
The Faculty of Science; UCPH, has supported the DIKU part of this course economically, thus DIKU PhD students must sign up here.

Click the search button to search Courses.

Course calendar
See which courses you can attend and when

Publication of new courses
All planned PhD courses at the PhD School are visible in the course catalogue. Courses are published regularly.