Du är inte inloggad på KTH så innehållet är inte anpassat efter dina val.
Topic: Representation Learning course - A broad overview Zoom Meeting
Time: This is a recurring meeting Mondays 8:30
Join Zoom Meeting
https://kth-se.zoom.us/j/62189868270
Meeting ID: 621 8986 8270
Project ideas:
https://docs.google.com/document/d/1qR5H2Htic1EJF-AeZNXtKjWA_RKxIgPv9OZKySw0wm4/edit?usp=sharing
This course has the goal to give a broad overview of the current state of the art in Representation learning.
We will tackle four topics ( disentanglement, generative models, graph representations learning, and Hierarchical representations) with an in-depth prepared lecture and a paper discussion session for each topic.
The course is done in students teach students manner and has the students prepare and hold a 2h lecture about a direction in representation learning.
The following session is then structured as a discussion where the most relevant paper(s) are discussed as well as open questions from the lecture collected.
The course will therefore run for 8 weeks and end with a final project.
See the introductory slides for more detail.
The course can give you 6 credits in Topics for Computer vision with Dani as an Examiner.
Lecture assignment:
Disentanglement: Michael Welle, Giovanni, Youssef Mohamed
Generative models: Alberta, Marco Moletta, Petra Poklukar
Graph representations learning: Simon Holk, Aniss Medbouhi, Alex
Hierarchical representations: Gustaf, Yonk Shi, Alfredo