Uni-Tübingen

attempto online

17.06.2021

University of Tübingen and SiMa.ai Collaborate to Accelerate Low-Power Artificial Intelligence at the Embedded Edge While Reducing AI Compute Carbon Footprint

SiMa.ai Technologies (San Jose, California) announced its joint collaboration with the Department for Computer Science at the University of Tübingen in the Hardware-Agnostic Artificial Intelligence for Embedded project. Through this important research project, the University of Tübingen is bringing together the academic community and top industry leaders to address and bridge the gap from high performance, high-energy embedded compute to high performance, low-power compute for embedded edge applications.

A major focus of the research project will be to identify game-changing AI solutions that offer ultra-low energy consumption along with software architecture and approaches that minimize the dependency on hardware. SiMa.ai, whose purpose-built MLSoC™ platform delivers high-performance machine learning at the lowest possible power, was selected to participate due, in part, to its software-centric architecture that is built on the Tensor Virtual Machine (TVM) open-source frontend. Software solutions that minimize hardware dependencies greatly reduce engineering development costs and aid in accelerating application migration, development, and deployment velocities. As part of this research project, the SiMa.ai software tool chain will run on the university’s vast database of open source and custom neural networks for an in-depth analysis of power and performance metrics. The results of the study will aid machine learning engineers in optimizing and running computations efficiently and effectively to get the most out of their hardware.

“SiMa.ai was recently named one of the world’s leading startups in green machine learning technology so we are thrilled to be partnering with them for this important project,” said Prof. Dr. Oliver Bringmann, Head of the Department of Computer Science and the Chair for Embedded Systems at the University of Tübingen. “Their software-first approach to machine learning acceleration will greatly help this initiative as well as our collective efforts to make low-powered AI offerings more available and easier to adopt.”

In the spirit of Cyber Valley, one of the largest AI research cooperations in Europe, this research project further supports Germany’s national goal and the European objective to reduce the overall CO2 footprint by enabling energy-efficient AI compute. Machine learning and the cloud infrastructure have an enormous carbon footprint and are becoming one of the largest consumers of energy on the planet. For machine learning and AI to scale in adoption at the edge, it is critical that attention be paid not only to performance, but also to power. SiMa.ai was chosen to partner with the Embedded Systems research group on this project because it is a key technology innovator in this area.

“Fostering the exchange between the research community and industry leaders is an important step in fueling innovation at the edge,” said Krishna Rangasayee, founder and CEO of SiMa.ai. “The University of Tübingen is a recognized center of excellence for artificial intelligence and SiMa.ai is very pleased to be working alongside them to further machine learning acceleration and scaling while simultaneously working towards our common goal of carbon footprint reduction.”

SiMa Technologies/Embedded Systems research group

Contact:

Prof. Dr. Oliver Bringmann
University of Tübingen
Embedded Systems research group
 +49 7071 29 77348
oliver.bringmannspam prevention@uni-tuebingen.de
http://www.embedded.uni-tuebingen.de

Zurück