The Electronic Visions group's research focuses on bridging the gap between biological information processing and conventional numerical computing. Our focus is on designing, operating and leveraging our custom neuromorphic hardware systems capable of emulating complex neural dynamics and plasticity directly in silicon. In conjunction with programmable embedded SIMD processors, this allows for the creation of flexible, highly efficient real-time learning and inference systems.
We are developing massively parallel, mixed-signal VLSI neuromorphic hardware and full-stack software to operate these systems and to provide a user-friendly experimental platform for neuromorphic computation. Furthermore, we are interested in translating concepts from machine learning, particularly deep neural networks, into the 'spiking' world to build energy-efficient and scalable systems.
The research group is led by Prof. Dr. Johannes Schemmel, Chair of Neuromorphic Computing Architectures at the Institute of Computer Engineering. We are located at the European Institute for Neuromorphic Computing (part of the Kirchhoff Institute for Physics), a state-of-the-art facility dedicated to the exploration of new paradigms of information processing based on the concept of 'Physical Computing'.
Our BrainScaleS-2 (bɹeɪnskeɪlz) hardware is currently offered via the EBRAINS Research Infrastructure, see ebrains.eu/nmc for details on the neuromorphic resources.
For further details also see group history.
HOME
