Generative AI - 2025 entry

Generative models are widely used in many subfields of AI, making Generative AI a rapidly evolving and transformative field. In this module, you will study the theoretical foundations, for example, the probabilistic foundations and learning algorithms for generative models, variational autoencoder (VAE), and generative adversarial networks (GANs). You will also study application areas that have benefitted from generative models. You will attend lecture sessions, complemented by lab sessions, allowing you to apply your knowledge through hands-on exercises. You will work with popular AI frameworks and generative AI libraries, gaining practical experience. This module is suitable for Computer Science, Mathematics and Engineering students and any students with experience in programming and deep learning. 

Co-requisite module: COMM113 Deep Learning

This module aims to provide you with a comprehensive understanding of the probabilistic foundations and learning algorithms of generative models, as well as some generative AI applications. You will study techniques for example, Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), autoregressive models, energy-based models, and diffusion models. You will learn theoretical principles and practical implementations using machine learning and deep learning frameworks and libraries, enabling you to design, train, and evaluate generative models effectively. Through hands-on exercises and assessments, you will develop the skills needed to develop and apply generative AI techniques in research and industry, prepare you for a career in this transformative field.

On successful completion of this module you should be able to:

Module Specific Skills and Knowledge:

1. Explain the foundations and advanced models of Generative AI, including for example model architectures and training techniques.

2. Implement various generative model solutions using deep learning frameworks, optimising for specific tasks, for example, image generation, language synthesis, or other creative outputs.

3. Evaluate the designed generative model solutions for limitations, strengths, improvements required, etc.

Discipline Specific Skills and Knowledge:

4. Formulate relevant real-world problems as generative AI problems, for example problems of data generation and multi-modal interactions.

5. Critically evaluate applications of generative AI to real world problems, including human and organisational impact.

Personal and Key Transferable/ Employment Skills and Knowledge:

6. Effectively communicate findings and evaluations drawn from research papers.

7. Demonstrate independent study and research skills through conducting projects.

• Goodfellow, I., Bengio, Y., Courville, A. and Bengio, Y., 2016. Deep learning (Vol. 1, No. 2). Cambridge: MIT press
• Bishop, C.M. and Bishop, H., 2023. Deep learning: Foundations and concepts. Springer Nature.
• Generative AI for Beginners Team, G., Anil, R., Borgeaud, S., Alayrac, J.B., Yu, J., Soricut, R., Schalkwyk, J., Dai, A.M., Hauth, A., Millican, K. and Silver, D., 2023. Gemini: a family of highly capable multimodal models. arXiv preprint arXiv:2312.11805.
• Rombach, R., Blattmann, A., Lorenz, D., Esser, P. and Ommer, B., 2022. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (pp. 10684-10695).

• ELE 

Reading list for this module: