Information and material for the course Generative AI at The Arctic University of Norway
The new version of the course is coming! The next edition will be in spring 2025.
The course will be streamed online and recorded. You can attend physically at the Tromsø (main place) and Bodø campuses.
Course code for this year:
- INF-3993-3, for master students
- INF 9301, singular courses code at PhD level
Canvas page of the course (you need to be registered to see this page)
The course covers Large Language Models (LLMs), Generative AI for images as well as applied aspects such as continuous integration in production systems. The goal of the course is to provide a master-level introduction to Generative AI, combined with hands-on experience about these new technologies. The students will practice current state-of-the-art open-source tools for generative AI. At the end of the course the students will understand the concepts behind generative AI and be able to program and interact with their own generative models. The course is for students passionate about AI and generative AI. Some emphasis will be put on creating AI tools, which could lead to possible startup ideas. Programming skills in Python are required. Familiarity with machine learning terminology is expected. Knowledge of deep learning is not mandatory but recommended.
You will learn about
- the transformer achitecture and Large Language Models, how they work and how to run them (small versions) locally on your computer, how to fine-tune them,
- LLMs and their limits, are they stochastic parrots?
- how you can combine LLMs with other tools for retrieving information in documents or making a recommender system, what RAG "Retrieval Augmented Generation" is,
- Generative AI for images and diffusion models, the principles and the practice with Huggingface and tools such as comfi.ui,
- LLM reasoning and multimodal,
- Ethics in AI,
- and more!
You can find more information on the Schedule in the file scheduleGenerativeAI2025.pdf. The schedule is preliminary and may be subject to small changes.
You will be evaluated on a team project you develop over the semester. This project will be a software or demo making use of the tools and concepts we see in class (see folder of the previous year for some examples). PhD students will have an additional task of presenting a scientific paper during one of the lectures.
You can register if you are:
- a master student at UiT
- a PhD student at any university in Norway
Master students: register on studentweb.
Instructions for PhD students (deadline for registering a PhD student: 1st of december): https://uit.no/utdanning/opptaksinfo#innhold_711494
Course leaders:
- Helge Fredriksen (Bodø campus), [email protected]
- Benjamin Ricaud (Tromsø campus), [email protected]
You can find information about previous years in the dedicated folders