Partnering with the EPSRC funds vital defence research
Dstl’s partnership with the Engineering and Physical Sciences Research Council (EPSRC) funds crucial defence research, including AI to boost intelligence analysis.
The Defence Science and Technology Laboratory (Dstl) works closely with the Engineering and Physical Sciences Research Council (EPSRC) to provide a future pipeline of highly qualified staff for priority defence areas.
Dstl and EPSRC have co-invested in a number of schemes, which include:
- MOD centres for doctoral training
- Dstl allocated doctoral landscape awards
- AI and Quantum hubs
- Future Compute
Richard Walters from the Alan Turing Institute explains how Dstl’s partnership with EPSRC funds crucial defence research, including AI to boost intelligence analysis.
I took part in an EPSRC-organised Sandpit; a unique research-funding mechanism that seeks both to incubate innovation and drive engagement between academia and defence.
The week-long event was effectively a sort of science and innovation ‘lock-in’, with a diverse group of 60 attendees spanning academics, Dstl and other UK government scientists, plus defence and national security problem-owners, focusing on the use of AI to support decision-making in national security and defence applications.
How AI supports decision-making in the military
Intelligence analysts must make high-consequence, defensible assessments, often using large amounts of complex and uncertain data to spot early signs of threats or hostile actions. They have to work quickly under pressure with limited resources, deciding which data to analyse first and whether to risk gathering additional intelligence.
More about the Sandpit
Richard leads a 2-year, £1.25 million programme called AI Intelligence Triage and Acquisition Support for Human-centred Analysis (AiTASHA). This will build explainable, defensible AI to complement, rather than replace, the work of intelligence analysts.
The project will improve the speed and confidence of intelligence analyst’s assessments. It will do this by recommending which existing data should be prioritised for human review, and which potential new data would be most useful to collect next.
This research involves a diverse team working closely with defence and national security partners, which includes:
- the Alan Turing Institute
- Warwick University
- Southampton University
- Heriot-Watt University
- Cardiff University
It’s one of 2 successful projects funded from the Sandpit.
Advantages of a Sandpit
The ‘lock‑in’ format made it quick and effective to filter high-quality research ideas in a very short period; over 1,500 hours of people’s time went into the week-long event.
The event setup also raised the quality and ambition of the projects; AiTASHA features not only deep learning and statistics research, but will also explore areas such as human factors and AI ethics as well.
Additionally, it was valuable having direct, extended access to government scientists and problem owners, which meant working together from day one to create projects with real‑world impact.
Read human-centred ways of working with AI in intelligence analysis biscuit book, or any of our other AI biscuit books.