COMP 6240 - Generative AI 3

This course is an advanced study of Generative Artificial Intelligence (GenAI), focusing on the deep learning architectures and probabilistic models that enable the creation of novel and realistic data across modalities (text, image, audio, and code). Designed for graduate students with a strong foundation in deep learning, the course provides both the mathematical theory and practical implementation skills necessary to work at the forefront of this rapidly evolving field. Students will delve into the theoretical underpinnings and practical training challenges of core generative models: Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), Transformer Models and its application in Large Language Models (LLMs), including pre-training strategies, prompting, and fine-tuning (e.g., LoRA), and Diffusion Models. Prerequisite(s): COMP 5480 Artificial Intelligence. Hybrid