Machine Learning

Machine learning – computer algorithms and systems that use data to learn, adapt their behaviour and make decisions – is one of the key disciplines that underpins artificial intelligence. 

Some of our most celebrated machine learning breakthroughs, such as support vector machines and kernel methods, began as theoretical models before finding their way into the tools and applications we use every day. 

Other innovations, such as  foundational models for medical imaging, which help us understand more about Alzheimer’s and dementia, lung disease, and eye conditions, have now been adapted for many other uses. 

This broad spectrum of activity, from abstract theory to problem-solving application, is the true source of our machine learning research strength.

How is UCL Computer Science’s machine learning research making a difference?

Generative AI leadership

To maintain the UK’s status as a global leader in AI research and innovation, the AI Hub in Generative Models, funded by UKRI, unites experts from academia and industry for a more strategic approach to generative AI research. Working together will allow us to find and fund nationwide solutions to the most complex technical and social challenges in machine learning.

Led by UCL, the generative AI hub combines the research strengths of Imperial College London, Cardiff, Cambridge, Oxford, Manchester, Edinburgh and the University of Surrey. Industry partners include IBM, BT, DeepMind and Cisco Systems.

Sustainable AI

Machine learning is a fundamental part of many of the technologies being developed to tackle the climate crisis. Creating detailed climate models based on both real-time and historic data helps us understand complex interactions between factors like energy consumption, fossil fuel production, extreme weather events and carbon emissions. 

Using these models, we can predict and recommend optimal courses of action for disaster response and recovery within a few hours, as well as policy recommendations and pathways to help manage the energy transition over the coming decades.

We’re also working to address the sustainability issues surrounding the energy consumption of foundation AI models. We’re reducing the carbon footprint of AI by improving the algorithms, with more efficient transformers and diffusion models. We’re also using machine learning to improve the stability and energy efficiency of nuclear fusion processes, increasing the amount of low-carbon energy available to meet demand.

Spinouts and startups

UCL Computer Science machine learning researchers operate across both industry and academic environments. Spinouts and startups play a huge role in helping us strengthen and nurture this cross-fertilisation, and helping us design algorithms which are useful and valuable to society.

Our Venturer programme streamlines the spinout process, with three established pathways that prioritise innovation, whether it originates from one of our academics, a prospective PhD student or an industry funder seeking to solve a problem.

AI companies cofounded by UCL Computer Science academics or students, including Synthesia, Matrix Mill, Satalia and Genie AI, have been collectively valued at more than £2 billion.

Healthcare

Machine learning’s capacity to analyse and identify patterns in huge datasets is helping us understand and treat disease in entirely new ways.

From the Nobel Prize-winning AlphaFold collaboration between UCL computational biologists and Google DeepMind, to virtual biopsies and robotic surgery that allow us to diagnose and operate on difficult-to-reach cancers with less physical trauma, machine learning is redefining what’s possible in healthcare and medicine.

What’s coming next?

Machine learning is now an inseparable part of academic research, across UCL Computer Science and beyond. As we move towards a future where we entrust more of our workloads, our responsibilities and ultimately, our hopes to artificial intelligence, our research in machine learning will help us ensure that these technologies can be relied upon.

Associated research groups and centres

Related MSc courses