Project Team wins Award for Energy Efficient Neuromorphic Hardware Design

Symbolbild zum Artikel. Der Link öffnet das Bild in einer großen Anzeige.

Creating an energy efficient chip, which can classify ECG signals into healthy and arrhythmical using neural networks, was the goal of the first challenge for disruptive innovations in energy-efficient AI hardware initiated by the German Federal Ministry of Education and Research (BMBF). The Project Team „Low-Power Low-Memory Low-Cost ECG signal analysis using ML algorithms (Lo3-ML)“ under the Leadership of Prof. Dr. Dietmar Fey, Dr. Marc Reichenbach (Chair of Computer Science FAU), Prof. Dr. Dr. Robert Weigel (Chair of Electronics FAU), Dr. Marco Breiling (Fraunhofer IIS Erlangen) and Prof. Amelie Hagelauer (Chair of Communications Electronics, University of Bayreuth) won the first price in the category “ASIC 130 nm”. Federal minister Anja Karliczek presented the award in a virtual award ceremony on March 18, 2021. The design combines a data flow driven architecture, non-volatile, multi-level RRAM memory and a division into two parts with a specifically adapted NN training algorithm. It achieves a very low power consumption and can be applied in a wearable use case.

Press releases in German language from FAU and Fraunhofer IIS can be found here and here.


A special thanks to all team members and all others involved in the project for the great collaboration:
Dr. Marc Reichenbach
Dr. Marco Breiling
Maen Mallah
Stefan Pechmann
Timo Mai
Peter Reichel
Daniel Reiser