Neuromorphic Computing
The human brain does not process information as bits (like classical computers), but rather uses impulses (spikes)—that is, short electrical signals. What matters is when such an impulse occurs—the exact timing conveys the actual information.
As a result, the brain can convey more and more complex information per signal than a digital bit, which only recognizes “0” or “1.” The time between spikes—or their pattern—contains significantly more information, making this method more efficient for certain tasks—such as recognizing patterns or sequences in real time. In technology, such concepts are replicated in “neuromorphic computing” through so-called spiking neural networks (SNNs).
“Our goal is to develop systems that mimic biological structures and processes—with minimal energy consumption and maximum responsiveness,” explains Dr. Bernhard A. Moser, Technology and Innovation Manager at the SCCH. “Applications include robotics, medical technology, and environmental monitoring, among others.”
Research Team Dispels Misconceptions
Despite the growing interest in neuromorphic computing (NC), some persistent misconceptions remain, such as the belief that it requires expensive, specialized hardware, that it is slower than traditional methods, or that it lacks industrial maturity. “We’re showing that many of these assumptions are outdated. First, NC offers enormous potential not only in terms of energy efficiency but also in terms of speed. Second, we’re demonstrating that NC is by no means limited to specialized analog hardware but can also be used effectively on standard hardware platforms,” emphasizes Dr. Thomas Buchegger, site manager of SAL in Linz.
A new study illustrates the versatility of neuromorphic computing—using an example from the field of medicine. Researchers from Johannes Kepler University (JKU), SAL, SCCH, and Graz University of Technology (TU Graz) collaborated to investigate how ECG data can be collected and processed more efficiently. In doing so, they were able to reduce the data volume by over 80 percent without compromising the quality of the analysis. This means less storage space is required while maintaining the same level of accuracy. This technology opens up great opportunities for the future of medical diagnostics and beyond.
Milestones of the Research Partnership
Since the start of their collaboration, the partners have achieved a series of groundbreaking successes that impressively underscore the international potential of “Made in Austria” neuromorphic computing:
- World Record in SNN Inference on Standard FPGAs
Using a new demonstrator, researchers have successfully shown how so-called spiking neural networks (SNNs)—a special type of artificial neural network—can be run exceptionally quickly on common FPGA systems. FPGA stands for “Field-Programmable Gate Array” and refers to a flexible, customizable computer chip that is frequently used in engineering.
“With our method, we can process over 2.5 million images per second—that’s more than 100 times faster than previous systems using the same hardware and data,” explains Associate Professor Dr. Michael Lunglmayr from the Institute for Signal Processing at JKU. “In addition, our system is very energy-efficient, achieving a rate of more than 3 million images per second per watt.” This breakthrough demonstrates how powerful and, at the same time, energy-efficient modern artificial intelligence can be implemented on off-the-shelf hardware.
A patent application has already been filed for a new invention for particularly energy-efficient data acquisition. It utilizes neuromorphic principles inspired by the functioning of the human brain.
- Austrian Workshop Series and Publications
With the SNNSys workshop series on neuromorphic computing and spiking neural networks (SNNs), the collaborating partners are providing important impetus to the Austrian research landscape. The workshops are organized as part of the AIROV conference in close collaboration with Graz University of Technology (TU Graz) and promote academic exchange at the highest level. Research results from this context have already been published multiple times in renowned journals such as the *Journal of Neuromorphic Computing and Engineering* as well as at leading conferences, such as the IEEE International Symposium on Circuits and Systems.
- Organization of an International Competition
In September 2025, the major IEEE Conference on Image Processing (ICIP) will take place in Alaska. Our research team is organizing an international competition under the motto “Low Power, High Speed.” The goal is to find the best solutions for particularly energy-efficient and fast AI.
“We are very much looking forward to the challenge. The strong response from the professional community shows just how much creativity and innovative spirit there is in this field,” says Moser.
About the Research Partners
Software Competence Center Hagenberg (SCCH)
The Software Competence Center Hagenberg (SCCH) is a non-university research center that has built a track record of excellence in application-oriented research in the fields of data science and software science for over 20 years. This focus enables the optimal implementation of projects in the fields of digitalization, Industry 4.0, and artificial intelligence. The SCCH sees itself as an interface between international research and the local economy and conducts outstanding research with its team of researchers. Currently, the SCCH employs approximately 130 staff members from 25 different countries. The SCCH is supported as a COMET Center—Competence Centers for Excellent Technologies—by the Federal Ministry of Transport, Innovation, and Technology (BMVIT), the Federal Ministry for Digitalization and Business Location (BMDW), and the State of Upper Austria.
www.scch.at
Johannes Kepler University Linz (JKU) – Institute for Signal Processing
The Institute for Signal Processing (ISP) at Johannes Kepler University Linz focuses on the algorithmic, architectural, and hardware-oriented aspects of signal processing systems. Areas of application include information and communication systems, high-frequency and baseband ICs, bio- and sensor signal processing, and automotive applications. Our vision is to work on both long-term basic research projects and short-term topics, the latter in close collaboration with industry partners.
Johannes Kepler University Linz (JKU) has been a university with strong regional roots and an international focus for nearly 60 years. It is the largest educational and research institution in Upper Austria, offering teaching and research in the fields of law, engineering, natural sciences, medicine, education, and social and economic sciences. Approximately 25,000 students and nearly 4,000 employees enjoy a modern infrastructure on the JKU campus with plenty of space for learning, research, and living.
Silicon Austria Labs (SAL)
Silicon Austria Labs (SAL) is Austria’s leading research center for electronics- and software-based systems (ESBS), which form the technological backbone of digitalization. At three locations in Graz, Villach, and Linz, SAL conducts cutting-edge research in the fields of microsystems, sensor systems, power electronics, intelligent wireless systems, and embedded systems to develop future-oriented solutions for industrial production, healthcare, energy, mobility, and security. The Intelligent Wireless Systems division focuses on developing reliable and secure communication systems for industrial environments. This enables broadband connections for mobile devices (e.g., smartphones, tablets, laptops), wireless machine-to-machine communication, and, increasingly, wireless sensor networks in industrial applications.
One of SAL’s flagship initiatives is Digineuron, a pioneering project aimed at developing highly efficient integrated circuits that implement AI in miniature formats. These chips mimic the structure and function of the human brain and use multilayer neural networks that communicate locally to minimize energy consumption. This opens up new possibilities for intelligent, low-power applications. SAL brings together key players from industry, academia, and research, combining valuable expertise and know-how to conduct collaborative, application-oriented research across the value chain. Collaborative projects are co-funded by SAL, enabling a quick and streamlined project launch.
www.silicon-austria-labs.com
