Programme A

Foundation Models

Fundamentals & Applications

Build from representation learning, model architectures and pre-training to multimodality, evaluation and transfer, then work through cases in finance, music and robotics.

June 2027
Online

/Four connected modules

I · Core

Core Theories & Frameworks

The common theoretical ideas and frameworks behind modern foundation models.

Topics

Representation learning
Transformers
Autoregressive modelling
Masked modelling
Diffusion models
Tokenisation
Scaling laws & pre-training

II · Advanced Topics

Architectures & Capabilities

The architectures and capabilities extending the core foundation-model paradigm.

Topics

Emerging architectures
Multimodality
Long-context learning
Retrieval & grounding
Memory
World models
Evaluation & transfer
Adaptation & post-training

III · Applied

Application Domains

Explore how foundation-model principles apply across selected domains and transfer between them.

Domains & transfer

Finance
Music & audio
Robotics
Cross-domain transfer

IV · Labs

Hands-On Cases

Three practical sessions for applying foundation-model principles across selected domains.

Coming soon. Lab topics will be announced.

Who it’s for

Researchers, engineers and technical leaders

People who want a strong, cross-domain foundation before specialising in an application field. A working knowledge of machine learning is expected.

What you’ll gain

A coherent foundation-model framework

Understand the main theoretical ideas, model families and training objectives behind foundation models.
Connect multimodality, long context, retrieval, memory and world models.
Understand how emerging architectures are evaluated, adapted and transferred.
Apply foundation-model principles across finance, music and robotics through focused labs.

/OxML 2027 registration

Find your programme.

Compare programmes and prices, then complete your registration.

View registration options

Loading…