From detecting creatures on the seafloor to demoing a digital twin of the UK energy system, the Turing’s research highlights of the past year showcase the diversity of our work with industry, academic and government partners.
These 10 research projects are profiled in the Turing's Annual Report 2022-23. They represent just a taste of last year’s successes – read the report for the full lowdown on all our activities.
With the recent release of our Institute strategy, which has a renewed scientific focus on the societal grand challenges of health, environment, and defence and national security, the next year will see us doubling down on interdisciplinary research and innovation as we continue to use data science and AI to change the world for the better.
1) Developing a more accurate measure of the nation’s health
We worked with the Office for National Statistics to carry out an in-depth statistical analysis of its Health Index, which generates annual scores for England’s health at national and local authority level. Our analysis ensures that the index is as accurate as possible, providing high-quality data that can be used to develop policies for improving people’s physical and mental wellbeing. Read more.
2) Revealing hidden details of the ocean floor
Studying the creatures living on the UK seafloor is essential to advancing our understanding of this fragile habitat. We teamed up with the government’s Centre for Environment, Fisheries and Aquaculture Science (Cefas) to develop an algorithm that automatically detects and classifies animals called sea pens – otherworldly structures that offer insights into the health of their seafloor ecosystems. Read more.
3) Demoing a digital twin of the UK energy system
Developing policies aimed at achieving net zero greenhouse gas emissions by 2050 is a difficult task, but a digital twin of the UK energy system could help. We led the technical development of a demonstrator for what such a digital twin might look like, providing the government with a proof of concept for a tool that could ultimately help to predict the long-term impact of energy policies. Read more.
4) Pioneering new approaches to verifying the fairness of AI models
AI models can save companies time and money by automating decision-making processes, but for companies to gain the trust of their employees, clients and the wider public, they need to be able to prove that their models have been developed in as fair a way as possible. We have created a unique methodology that uses a cryptographic technique to confidentially verify a model’s fairness in less than two minutes. Read more.
5) Informing UK security policy through evidence-based research
Our Centre for Emerging Technology and Security (CETaS) is conducting research and analysis across a range of security-related topics, from open-source intelligence to disinformation to AI voice-cloning technology. By engaging with senior policy makers, CETaS is directly informing UK government understanding of the threats posed by, and available responses to, nascent technologies. Read more.
6) Enabling researchers to work safely with sensitive data
Over the past five years, we have developed a digital, cloud-based infrastructure that allows our researchers to use trusted research environments (TREs) to work safely with sensitive data in domains like health and finance. This year, we made the project open source – a major milestone that allows other institutes to freely adopt and adapt our code and documentation to deploy and manage a TRE of their own. Read more.
7) Harnessing the power of data in cities around the world
Local authorities are increasingly recognising that, in order to improve the lives of city dwellers, they need detailed data on a city’s buildings – the homes, offices, shops and schools where we spend most of our lives. Our Colouring Cities project is making this possible through the creation of interactive maps that collate, store and visualise buildings data for cities including London, Athens and Sydney. Read more.
8) Championing next-generation encryption techniques for the finance sector
Tackling financial crime such as money laundering requires banks to share data with each other and pool their intelligence, but privacy requirements can make this difficult. A cutting-edge technique called homomorphic encryption (HE) promises to make it easier for banks to securely share data. We have been working with HSBC to carry out an in-depth review of HE’s current usage in the finance sector. Read more.
9) Understanding online abuse of Premier League footballers
We published a report in collaboration with Ofcom showing that seven out of ten Premier League footballers face abuse on Twitter. The analysis was carried out by the Turing’s Online Harms Observatory – a tool that uses machine learning to automatically detect abusive online content. We aim to use this tool to inform future policy around online harms and help make the internet a safer place for all. Read more.
10) Powering population models with synthetic data
A new Turing-developed tool – the Synthetic Population Catalyst – is making it easier for researchers to access synthetic data for human population models. The anonymised data generated by the tool includes variables related to personal health, household type, employment status and social interactions, and is already being used to model the individual health impacts of climate change-related extreme heat. Read more.
Read the Turing's Annual Report 2022-23 here.
