Course on Large Language Models

Transformers and large language models (LLMs) have quietly become the engine behind modern language AI - but most people only see the output, not the machinery.

This 6-lecture, 12-hour technical lecture series at the MDIA explains, over 6 evenings, how LLMs actually work: word embeddings, the Transformer block, and the Query-Key-Value self-attention mechanism that powers GPT-style modes.

It's not a "how to use Chat GPT" course; it is a deep, engineer-and-programmer-friendly dive into the architecture and training of these huge models, why hallucinations happen, and what today's models do (and don't) imply about AGI.

Apply if you want the real internals, not the hype.

Registration is free and places are allocated on a first-come, first-served basis.

More information is available online.

Registration is available online.

Lectures will be delivered (in English) by Prof. John Abela. John Abela is a professor in the Faculty of ICT. His main research interests are AI, Machine Learning (including Deep Learning), Machine Vision, computational complexity and NP-completeness, search and optimisation algorithms, Large Language Models (LLMs), and Quantum Computing.

Additional information is available by email.

All lecture slides will be made available to registered attendees.