National initiative for the next generation of AI

Sensor and computer technologies that mimic the brain’s structure and function, known as neuromorphic technology, are paving the way for a new generation of AI that is significantly more energy efficient and adapted to machines, equipment and sensor systems. Vinnova has now granted a project coordinated by Luleå University of Technology for national collaboration to prepare a cluster of excellence.

"Internationally, there are clear roadmaps for the development of neuromorphic technology, and we need to create the same in Sweden. That is why it is very positive that we can now develop a common vision and strategic plan to strengthen Swedish industry through neuromorphic technology. To keep up with the rapid global development, we need to come together nationally and collaborate across disciplines towards long-term common goals", says Fredrik Sandin, Professor of Machine Learning at Luleå University of Technology and scientific leader for the feasibility study.

Behind the feasibility study Shaping the Future with Neuromorphic Technology (NEUTEC), in addition to Luleå University of Technology, are ABB Robotics, Ericsson AB, FOI, Frontgrade Gaisler AB, KTH, Linköping University, Lund University, RISE, SAAB AB, Terranet AB, Uppsala University and Örebro University. The study aims to develop a common vision and strategic plan to create a Swedish cluster of excellence for next-generation intelligent systems, and welcomes more universities and companies to join the open workshops that will be arranged in the coming months.

AI with minimal energy

AI creates many opportunities, but today’s computing architectures are extremely energy demanding, and AI development depends on increasingly large and energy-intensive data centers. Neuromorphic technology imitates the brain’s way of processing information, with circuits that learn and adapt like neurons and synapses. This enables embedded AI with minimal energy use and very fast response times.

– Neuromorphic technology is about the next generation of AI, taking new steps in the development of artificial intelligence by mimicking the brain’s structure and function. Neuromorphic sensors and intelligent systems can become thousands of times more energy efficient than today’s AI, and can therefore be integrated into equipment for much faster response times. Sensitive information can also be processed locally, which gives higher integrity, says Fredrik Sandin.

The technology is particularly relevant in complex and extreme environments, for example where cables should be avoided or where sensitive data should not leave the user’s equipment.

A previous Vinnova-funded feasibility study coordinated by Luleå University of Technology investigated the conditions for a Swedish innovation platform for neuromorphic technology. It showed that Sweden has strong research groups and committed companies but lacks national research programs and a coordinated strategy.

For more information, please contact Professor Fredrik Sandin or Project Manager Sabine Mayer.