ESGI hackathon challenges researchers to build code for artificial pancreas

6 Jul 2016

Shourjya Sanyal, CEO of ThinkBio Solution. Image via Vincent Hoban/UCD

When it comes to solving some of the major engineering and software problems that exist within major industries, it’s sometimes worth stepping back and getting a fresh perspective, as is the case at this week’s European Study Group with Industry (ESGI) event.

The ESGI event has established itself as one of the largest industry and academia collaborative efforts in the world, having first been launched in 1968 at Oxford University as a week-long problem-solving workshop.

The reasoning behind its creation is simple, in that, in some instances, a company reaches a brick wall of problem-solving that requires someone from outside to brainstorm and come up with a solution that might be right under their nose.

Future Human

With the first Irish event having been hosted back in 2008 by the Mathematics Applications Consortium for Science an Industry (MACSI) at the University of Limerick, this year’s event sees seven companies – five of which are Irish – collaborating with researchers from undergraduate level right up to senior professors with aims of solving some very complicated mathematical and statistical problems, much like a hackathon.

Solving really random problems

Organised by University College Dublin’s (UCD) School of Mathematics and Statistics, the event has attracted more than 70 researchers, not just from Ireland, but from a lot of far-flung places, too, including Australia, the United Arab Emirates and Israel, who are all eager to sink their teeth into some challenges.

Following a series of presentations by the companies, the researchers then break into teams and will spend the rest of the week tackling their chosen problem.

“It’s not everyday you get a really random problem [as a researcher],” said Dr Aoibhinn Ní Shúilleabháin who is one of the event’s organisers, in conversation with

“If you’ve got your special field, you’ll tend to just work on that. So, for example, if your background is in theoretical fluid dynamics, you mightn’t hear about this problem from Wupperman steel that there are defects that happen during the steel process.”

Taking this Wupperman Steel example, researchers are being posed the problem of the ‘banana’ defect found in steel production.

During the process, steel is coated with thick layers of zinc to prevent the occurrence of rust, however, banana-like defects occur near the edge of the strip where the zinc layer seems to creep upwards and sag again.

Mike Linnane

Mike Linnane, director of Evolve Technologies. Image via Vincent Hoban/UCD

Making an effective artificial pancreas

Another problem that researchers are being asked to solve is of particular interest to Dr Ní Shúilleabháin. It comes from the Irish start-up Evolve Technologies, which is looking to develop an accurate and trustworthy algorithm where an artificial pancreas can be used to supply the correct amount of insulin to a diabetes patient on a 24-hour basis.

Currently, type-1 diabetics are required to self-inject themselves with insulin four times each day, but with an artificial pancreas with an effective algorithm it could release sufficient insulin hundreds of times each day, much like a natural pancreas.

“They aren’t problems that we tend to hear about for mathematicians, but it could have real impact on the world and people’s daily lives as well,” Ní Shúilleabháin said.

Of particular focus this year for companies appears to be solving big data questions, with two companies – Think BioSolution and LittleData – putting problems to the international researchers.

Think BioSolution – founded by two UCD PhD graduates – is in the midst of developing technology that would allow a smartphone camera to give you health-related information.

For example, its technology allows a camera to read your heart and respiratory rate from a video of your face, but now they are seeking help to develop an algorithm that will help minimise the storage costs and data-processing time.


LittleData founder, Edward Upton. Image via Vincent Hoban/UCD

Irish start-ups involved

Aside from giving companies a way of recruiting researchers to solve their engineering and software problems, Ní Shúilleabháin said this event is a fantastic way for researchers to forge new careers by dealing directly with industry.

“Academia would always get a lot of critique that everything they do is in their ivory tower, while industry perhaps would get critiqued that everything they’re doing isn’t research-based,” she said.

“Here what we have is a perfect opportunity for those two sides to come together.”

Three of the companies taking part in EGSI are Irish start-ups – Ashgrove Anaytics, Think BioSolution and Evolve Technologies – which, Ní Shúilleabháin said, is something to celebrate.

“They’ve come from an idea that somebody got as a postgraduate, and now they’re building an industry and employing people while bringing in the research element.

“I think that’s really something to celebrate: that we have that entrepreneurship and that high-quality research where it all comes together.”

To see a full list of the companies and problems presented to the researchers, you can visit the ESGI microsite.

Colm Gorey was a senior journalist with Silicon Republic