Defence agency sets Data Science Challenge

Dstl joins with security agencies in calling on data scientists in step towards creating an ‘innovation ecosystem’ to solve problems in the sector

The Defence Science and Technology Laboratory (Dstl) has made a push towards bringing more data scientists into dealing with defence and security issues with the launch of a Data Science Challenge.

Data Science Challenge logoThe executive agency of the Ministry of Defence (MoD), which supports research into science and technology in the sector, has invited participants to come up with solutions for two problems with the lure of a total prize fund of £40,000.

One deals with detecting and classifying vehicles such as buses, cars and motorcycles from aerial imagery, aimed at dramatically improving how quickly they can be identified and assessed.

The other is focused on classifying crisis report documents. Analysing data in documents such as media reports can provide a better understanding of a potential crisis or a specific theme such as terrorism, and the challenge asks participants to predict topic tags for their classification.

Both challenges involve mining large unstructured datasets to extract useful information.

The competition is part of the Defence Innovation Initiative, which is aimed at building an open innovation ecosystem of individuals, academic institutions and private companies to develop new approaches to complex problems in defence.

Harnessing talent

Minister for Defence Procurement, Harriett Baldwin, said: “Our Innovation Initiative is about harnessing diverse and talented individuals from business, academia, and beyond to keep the UK ahead of our adversaries.

“In this latest challenge, supported by our £800 million Innovation Fund, we are calling on experts to develop the latest technology to crunch big data and identify the solutions that will keep us safe.”

James Srinivasan, a principal data scientist at Dstl, added: “By analysing complex, evolving information, data science can provide invaluable insight that informs how we can best respond to events.

“There is real talent out there and we want to encourage the curious to experiment and learn. We are determined to push the boundaries of what can be done, and to keep striving to always be better. This is why we are launching the Data Science Challenge.”

The Dstl is working on the challenge with BAE Systems, Capgemini UK and Roke Manor Research, and it is co-sponsored by the Government Office for Science, the Secret Intelligence Service and MI5.

The first prize will be £20,000, the second £12,000 and the third £8,000.

Image from GOV.UK, Open Government Licence v3.0