Energy-efficient AI

Researchers Bojian Yin and Sander Bohté at Amsterdam’s Centrum Wiskunde en Informatica (CWI, Centre for mathematics and computer sciences) together with a colleague in Eindhoven have achieved a mathematical breakthrough that will allow the development of energy-efficient AI to detect heart defects.

The solution means that special chips which are suitable for this AI model can recognise speech, gestures and electrocardiograms – the electric signals produced by the heart each time it beats – at a factor of 20 to 1,000 times more efficiently than traditional AI techniques.

Over the past decade, AI has gained more and more everyday applications, including for recognising images and spoken word. This is done with deep neural networks, which mimic the way the human brain processes information. For mobile applications, however, running current AI models often costs too much energy. Developing low-power AI has therefore become increasingly important. 

The energy gain from using this new algorithm on a special neuromorphic chip, means that, for detecting heart defects, an implanted ECG-recording chip will run for a year on a single battery.

Living lab for AI

CWI is part of the public-private collaboration AI Technology for People, backed by the City of Amsterdam, which encourages the application of human-centred AI technology that positively impacts society and citizens.

The Dutch capital has been recognised as one of the most AI-ready cities in the world and Amsterdam is proving to be an effective living lab for AI solutions. Dedicated hubs such as the Innovation Centre for Artificial Intelligence hosts labs researching uses for AI in areas as diverse as cancer care, cultural research, promoting equality and education.

Read more business news from around Amsterdam.