Leads: Lourdes Agapito (UCL), Yukun Lai (Cardiff), Pasquale Minervini (Edinburgh) 

Generative AI relies on human contributions, and we aim to support open-source model development through partnerships with key providers. By collaborating with them, we can combine UK research expertise and resources to create high-performance, open-source models and tools that meet societal needs.

Leads: Miguel Rodrigues (UCL) & Jack Stilgoe (UCL)

We will engage in policy discussions and collaborate with the UKRI programme on 'Enabling Responsible AI,' focusing on data contributors' rights, responsibilities in case of AI malfunctions, and using synthetic data to assess fairness.

Leads: David Barber (UCL) & Michael Wooldridge (Oxford)

With the global market for workplace automation expected to grow, our goal is to boost productivity by helping people perform tasks more efficiently and reliably.

Leads: Magnus Rattray, Jason Hartford (Manchester) 

Generative models have potential for analysing large, high-resolution biological datasets. We will explore a range of possible applications including, single-cell and spatial omics, microscopy imaging, and biological sequences such as DNA, RNA and proteins

Lead: Jose Miguel Hernandez Lobato (Cambridge)

This group focuses on advancing research where deep generative models meet molecular science. By combining machine learning with computational chemistry, we aim to revolutionise molecular design, enhance therapeutic and material properties, and uncover new insights into how molecules interact.

Leads: Aldo Faisal (Imperial), David Barber (UCL) 

We want to combine the deep technical competence of our Hub, with the domain expertise and data access of more applied Hubs. Our Hub has the great strength of being able to create new Generative AI algorithms that can be tailored to specific healthcare problems.

Lead: Arthur Gretton (UCL), Siddharth Narayanaswamy (Edinburgh)

Generative AI capabilities stem from recent theory and methodological breakthroughs in models like diffusion models, large auto-regressive models, and efficient state-space models. We will form a cross-Hub group to stay updated on the latest developments in these models and contribute to them.