Scientists are using quantum physics to bring AI to remote areas
James Dborin first became interested in low-memory machine learning when he realised this growing academic trend could help address some of the world’s most critical problems.
In the world’s remote areas and in outer space, using AI-powered tools can be a challenge. Running large machine learning models means relying on big computers and a stable internet connection, making it difficult to exploit the full potential of AI in situations where it could make a difference.
Enter TyTN. “Our technology allows you to deploy more powerful machine learning models on smaller computers with less memory,” Dborin says. “We move you away from faraway servers towards deploying edge devices that give you faster speeds and less reliance on an internet connection.”
In turn, this technology makes it easier to use AI in a wider range of places, such as in connection to satellites to detect natural disasters in real time and send out faster alerts, in agricultural tech or as part of conservation efforts to identify poaching in real time.
By relying on smaller devices, TyTN has also found a way to reduce costs at a time when the world is facing a severe chip shortage. “This is partially thanks to things like Bitcoin mining, which has caused a massive uptick in the price of silicon-based products, and so right now there’s a lot of value in being able to do what you want to do on cheaper devices,” Dborin says.
The idea for this solution came from the three founders’ background and interest in quantum physics — “where the problems are big and you need really big computers, in fact bigger than we could ever possibly build a computer, and you need to find ways to shrink down those problems whilst retaining their essence,” Dborin says.
At Conception X, the team hopes to commercialise the idea with the support of experts in their field, corporate partners, business and technology coaches.
Through the programme, TyTN was recently accepted onto Deloitte’s GRAVITY Challenge, which brings together corporates, entrepreneurs and universities to build solutions for some of the world’s most pressing problems through space technology. In particular, the team’s research will be integrated in a conservation project headed by the UK Hydrographic Office to optimise satellite and environmental data with the goal of minimising vessel collisions with marine mammals.
“We’re really excited,” Dborin says. “It’s a worthwhile cause but also a really good use case for our technology.”
Despite being very early stage, Dborin and team are currently focusing on finding a commercial use case to demonstrate the potential of their solution.
“Ultimately, we want to make a product that could be used by edge developers across a wide range of industries,” Dborin says. “My hope for five years would be to see this technology in the hands of a number of companies doing cool stuff and making a difference.”