Neuromorphic electronics

Sustainable electronics based on the human brain

What can we learn from our brains? Research has shown, for example, that our brains work much more efficiently than computers in many everyday tasks. By processing information in the same way as our brain, computer chips can use less energy. Scientists at the Center for Brain-Inspired NanoSystems (BRAINS) are working on electronics that are based on the way our brain works.

An example: you take a nice picture of your dog. The software on your phone 'sees' and recognises it as a dog. Before the software can distinguish dogs from other animals and objects, it has to be calibrated with tens of thousands of sample dog photos. This process is called machine-learning. This learning process takes a lot of time but, above all, a lot of energy. Even though the software only has to be tuned once, recognising your dog (the so-called 'inference' step) continues to take a lot of energy, because it still requires many calculation steps with a large claim on the device’s memory. In today's computers, the execution of these calculation steps, as well as the memory, are physically separated. The constant exchange of information between the memory and the processor slows down and is energy-inefficient. In the brain, however, information is processed and stored in the same place, which is in a communicating network of nerve cells (neurons). This makes recognising your dog a piece of cake for your brain. 

The brain as inspiration for electronics

The current generation of computers consumes a lot of energy. Within BRAINS, we are working on electronics that are based on the way our brains work. Scientists are working on completely new building blocks, consisting of materials that can behave like networks of neurons in our brain. The human brain as a source of inspiration not only makes for 'smarter' and more powerful computers but also, for example, for data centres that require a lot less energy.