Technical-embedded AI


The next level within AI-technology is bringing the artificial intelligence closer to the application. Machine Learning often leans on enormous amounts of data and transferring these raw data from the sensor to a central server far away can prove to be sub-optimal.

The technology will be more efficient if parts of this processing and adaptive learning already take place locally: close to, or even within the sensor. This asks for a thorough knowledge of not only the physical properties of these sensory data but also for an ingenious way of designing the underlying hardware. Special algorithms and architectures need to be designed to bring the cyber aspects of the system in concordance with the dedicated hardware, taking the overall context in mind.

This integrated view on Technical-embedded AI needs an equally multi-disciplinary approach: physicists, mathematicians, engineers and social scientists all working together to ensure that the smartness built in the devices is both efficient, robust and secure. Since long, this integrated and inclusive approach forms the core business of the University of Twente, founded as a university for the region and thus geared towards

Technical-embedded AI will find its application in almost every technical and societal domain, including health, smart robots and smart sensing for sustainable development goals.