Postgraduate / Master's · Artificial Intelligence
Generative AI
What is it
Postgraduate study of generative AI covers the mathematical foundations of generative modelling in depth — the probabilistic theory behind GANs, variational autoencoders, and diffusion models, along with open research questions around evaluation, controllability, and the theoretical limits of generated content quality.
Why it matters
Generative AI is one of the fastest-moving research areas in the field, and postgraduate study here focuses on the open problems — reliable evaluation metrics, mitigating failure modes, and understanding the theoretical properties of these models — that industry practitioners rarely need to engage with directly.
Exam tip
When comparing generative model families in a research context, be precise about what each is actually optimising (e.g. an adversarial objective versus a variational lower bound versus a denoising objective) — the training objective explains most of the practical differences in sample quality, diversity, and training stability.
Related topics
Want help mastering Generative AI?
Tell us about the student's goals and confidence — we'll design a personalised plan.