Advanced topics in Deep Learning

The course will cover the most recent developments in deep learning, which includes: 
Transformers with a focus on vision and attention. Multimodal Large Language models with a special focus on how information from different modalities are combined as well as on reasoning and transference of knowledge between modalities. Image generation with a focus on the latest GAN and Diffusion models. Graph Neural Networks with a focus on causality and other applications. After a lecture on the topics, students will present a recent paper on the area. In a session, two related papers are presented by two different students which will follow by a discussion comparing and analyzing both papers in depth lead by the presenters. Students will be evaluated in the paper presentation, discussion and reports/demos.