DATA 622 - Generative AI

(3)

This course offers an in-depth exploration of Generative Artificial Intelligence (AI), a branch of AI focused on creating models that can generate new content, such as images, text, and sounds, mimicking human-like creativity. The curriculum is designed to provide you with a deep background in the mathematics and computational techniques on which generative approaches are predicated, including but not limited to Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), and Transformer models. Throughout the course, students will: 1. Study the functional frameworks which underpin generative models: You will understand, and be able to derive as may be appropriate, equations inter-related with generative approaches (drawn from Bayesian studies and theory regarding diffusion models, variational autoencoders, and transformers, among others). 2. Explore Different Generative Models: Learn about various generative architectures, including GANs, VAEs, and transformers, understanding their unique characteristics and applications. 3. Apply Generative AI in Practice: Engage in hands-on projects that involve training, tuning, and deploying generative models to produce creative content across different domains, such as art, music, and natural language generation. 4. Implement your own class of Generative AI: Building on existing models, you will implement a new architecture for an application area using Generative AI. Programming assignments and projects will be carried out in Python, leveraging popular machine learning libraries like TensorFlow and PyTorch. By the end of this course, students will be proficient in designing and implementing generative AI models, capable of producing innovative and complex outputs that reflect aspects of human creativity. Cross-listed with DATA 446