Alan Turing Institute provides AI startups with opportunity to collaborate with researchers

Innovation centre Digital Catapult is partnering with the Alan Turing Institute to support nine London-based artificial intelligence (AI)-focused startups to develop their technologies further.

Participants will be invited to take part in the institute’s five-day study groups, which take place every few months, where they will be invited to present real-world problems to researchers, who then work with them to create a product to solve them.

The agreement with Turing will allow the startups to be involved in any study group from December 2018 to March 2020, which normally cost more than £20,000 to enter.

The Ministry of Defence agency and the Defence Science and Technology Laboratory previously used the event to find a machine learning product which improved aeroplane data analysis. It said with the help of the technology, the chances of a bug emerging in the system that could cause a major problem to the aircraft would be reduced.  

The Data Study Groups’ director, Sebastian Vollmer, said the event is a “collaborative community of motivated data science experts [which] team up to jointly create a seed for follow-up research and translational activities”.

“Each Data Study Group comprises a diverse mix of organisations at different stages of development. Startups and scaleups will come away from this experience with a fresh perspective, having had the opportunity to trial new cutting-edge research methods.”

He added the researchers in the study groups “range from PhD students to early career researchers to leading academics in their field”.

Digital Catapult’s CEO, Jeremy Silver, said: “The UK has the right ingredients to become a global leader in the development of AI and machine learning, and this partnership is a brilliant way of demonstrating how startups and scaleups can apply their ideas to real-world problems.” 

The partnership is another strand to Digital Catapult’s Machine Intelligence programme, which aims to drive growth in science and data startups.

The organisation has also been running a startup accelerator programme since January, selecting a new cohort every six weeks and providing them with access to compute resources.

After the initial announcement, Digital Catapult’s CTO, Marko Balabanovic, cited this as one of the four main barriers to growth for AI startups.

“You need a staggering amount of compute power to run this kind of stuff. You’re looking at these little startups having to spend 10 of thousands of pounds on cloud compute to do one training run,” he said.

Other challenges these companies face is a shortage of employees with the right skills, getting access to the required data and the ethical dilemmas of AI.