Campus Groningen is enriched with a new, unique institute for excellent research on the new generation of computers: CogniGron. Our current computer technology is pushing against its boundaries. New techniques with different materials are needed for a new generation of computer technology that uses less energy and processes data faster. Thanks to a gift to the Ubbo Emmius Fund, material scientists, mathematicians, computer scientists and artificial intelligence researchers will be able to work at the UG CogniGron research centre for the next seven years. The centre has been officially opened on 26 October 2018.
Developing 'smart' materials
The goal of CogniGron is to work with a multidisciplinary team from different areas of application in order to develop new, ‘smart’ materials, explains academic director Professor Beatriz Noheda. The materials will be put to use in computer systems that are inspired by our brains: ‘The current computers can already calculate better than our brains, but they are less good at recognizing patterns and assessing complex situations. This is why Google, for example, needs energy-devouring super-computers to translate speech to text, or to recognize images.’
The new generation of computers that CogniGron will work on are therefore not purely digital, like the current micro-electronics. Instead, the researchers will be looking at the neurons in our brains. These calculate all kinds of different values that depend on the input they get from their neighbours. This input can make them more sensitive or less sensitive. With these new computers, the classic division between memory and transistor disappears – which, in principle, makes them faster and more thrifty.
Bringing expertise together
To build this new sort of electronics, new materials are needed. These include, for example, a ‘memory’ that remembers previous input. The UG Zernike Institute for Advanced Materials (ZIAM) already has decades of experience in researching new materials. But models are also needed to optimally use these materials in circuits, as well as algorithms that can operate them. This is why the expertise of researchers from another UG institute is needed: the Bernoulli Institute for Mathematics, Computer Science and Artificial Intelligence. This is unique to CogniGron – in many places, research for these neuron-inspired computers is done from just one discipline.
What will this finally deliver? Professor Noheda: ‘In the coming years, we want to prove that new materials can indeed simulate a number of favourable properties of our brains, and that we can build complex systems that lay the foundations for a new type of computer.’
12 news professors
CogniGron is starting up with around 30 PhD students and a good 12 newly recruited professors. In addition, researchers who already work at the UG will also partially work within this research centre. These mostly comprise researchers from the ZIAM and the Bernoulli Institute. The researchers at CogniGron simply work within these research institutes. This way, synergy will form between the institutes and the research centre. As well as staff members, CogniGron has a budget of around 5 million euros for the infrastructure that is needed to make and study new materials.