Course description:

This seminar will cover the latest research in generative AI, with a focus on generative models, such as variational autoencoders (VAEs), generative adversarial networks (GANs) and diffusion models (DDPM, or recent modifications LoRA, IP adapters, SORA, …).

The seminar will be based on recent papers and will cover a wide range of topics, including the theory of generative models, training and evaluation of GANs, and applications of generative models in computer vision, natural language processing, and other areas.

Each student will be required to present a paper, write a summary of the paper, and participate in discussions of the papers presented by other students.

Prerequisites:

Basic knowledge of machine learning and deep learning (e.g., completion of the Neural Networks course). Basic knowledge of probability theory and statistics (e.g., completion of the Probability and Statistics course). Basic knowledge of linear algebra (e.g., completion of the Linear Algebra course). The seminar will be held in English (or Polish if all participants prefer).