Summer School 2024

Introduction:

Modern deep learning has had tremendous success in applying complex neural networks to problems from a wide range of disciplines, such as computer vision and protein folding. Geometric deep learning deals with incorporating symmetries into deep learning architectures. A symmetry of features is a transformation that is guaranteed not to change the labels. Symmetries are ubiquitous in many machine learning tasks. For example, in computer vision the object category is unchanged by shifts, so shifts are symmetries in the problem of visual object classification. In computational chemistry, the task of predicting properties of molecules independently of their orientation in space requires rotational invariance. This course gives and understanding of the theoretical basis underlying geometric deep learning. Furthermore, the course includes implementation of geometric components and as well as applying geometric deep learning on real-world data.

Learning objectives:

Upon completion of the course the student be able to

- understand the basic principles of geometric deep learning
- implement geometric deep learning algorithms
- compare various approaches in geometric deep learning
- read and critically assess geometric deep learning papers
- apply and evaluate geometric deep learning methods on real data sets

Course Plan:

Tentative Course schedule (the schedule may adjust in terms of topics and presenters).

Practical information:

Dates: 10-14 June 2024

Registration Deadline: 31 March 2024

Maximum Intake: 20 Candidates (Priority given to PhD candidates)

Total credits: 3 ECTS

Venue: H1 022, Floor 1, Eilert Sundts hus, Fakultet for Samfunnsvitenskap, Kristiansand Campus

Registration is closed 

Introduction:

Welcome to the "Secure and Robust AI Model Development (DAT945)" course. This course will explore theoretical frameworks and practical implementations, utilizing Jupyter Notebooks to enhance AI models against security threats, ensure robustness, and preserve privacy. Participants will develop a profound understanding of advanced techniques such as federated learning, homomorphic encryption, adversarial machine learning, and uncertainty quantification. The course emphasizes hands-on experience with Python, incorporating libraries like Keras/TensorFlow, uncertainty_wizard, PyFHEL, and Flower. Two assignments are planned throughout the course: a pre-assignment to assess participants' baseline knowledge and a post-assignment to evaluate their proficiency after training. Additionally, students will present their final projects, providing a comprehensive understanding of the practical applications of the acquired knowledge.

Course Plan:

Tentative Course schedule (the schedule may adjust in terms of topics and presenters).

Practical information:

Dates: 10-12 June 2024

Total credits: 5 ECTS

Registration Deadline: 31 March 2024

Venue: HU 062, Floor U, Eilert Sundts hus, Fakultet for Samfunnsvitenskap, Kristiansand Campus

Registration is closed