SynSense and INI propose a new approach for robust deployment of AI applications on mixed-signal neuromorphic processors
Researchers at SynSense and at the Institute of Neuroinformatics (INI), University of Zurich and ETH Zurich proposed a new approach for building robust low-power machine learning applications, in new work published in Scientific Reports.
Supervised training of spiking neural networks for robust deployment on mixed-signal neuromorphic processors: https://www.nature.com/articles/s41598-021-02779-x
The push for low-power AI on edge devices has lead to recent excitement for novel compute-in-memory architectures and mixed-signal analog/digital asynchronous computation. This approach is known as “Neuromorphic Intelligence” due to its inspiration from the biological nervous system.
The push for low-power AI on edge devices has lead to recent excitement for novel compute-in-memory architectures and mixed-signal analog/digital asynchronous computation. This approach is known as “Neuromorphic Intelligence” due to its inspiration from the biological nervous system.
In contrast to common digital information processing, neuromorphic devices use analog computing elements to achieve extreme power efficiency. Analog devices can have variation from circuit to circuit and chip to chip, which could reduce the accuracy of AI applications when deployed at scale, if not managed appropriately.
In this work SynSense and INI researchers propose a new way to build applications for asynchronous neuromorphic computing devices that overcomes this issue. “We used a network architecture designed to exploit sparse communication in mixed-signal neuromorphic devices, that protects against any noise and errors that can occur during operation,” said Dr. Dylan Muir, senior author on the paper.
The new approach is designed specifically for advanced mixed-signal devices from SynSense, but can be used to solve similar problems for other neuromorphic hardware in the research community. SynSense will use the method to build applications for future advanced devices that will exploit the benefits of both analog and digital circuits for ultra low-power AI edge-computing systems.
“Mixed-signal neuromorphic devices promise a huge additional power saving for edge and IoT devices,” said Prof. Giacomo Indiveri, SynSense co-founder, and co-author of the work. “This new work makes deploying AI applications on mixed-signal devices much more robust.”
About SynSense
SynSense is a leading-edge neuromorphic computing company. It provides dedicated mixed-signal neuromorphic processors which overcome the limitations of legacy von Neumann computers to provide an unprecedented combination of ultra-low power consumption and low-latency performance. SynSense has a unique technological edge and IP portfolio that comes from over 20 years of experience in mixed-signal neural processor design, advanced neural routing architectures, and neural algorithms.
To learn more about SynSense, please visit: www.synsense.ai
About Prophesee
Prophesee is the inventor of the world’s most advanced neuromorphic vision systems.
The company developed a breakthrough Event-Based Vision approach to machine vision. This new vision category allows for significant reductions of power, latency and data processing requirements to reveal what was invisible to traditional frame-based sensors until now. Prophesee’s patented Metavision® sensors and algorithms mimic how the human eye and brain work to dramatically improve efficiency in areas such as autonomous vehicles, industrial automation, IoT, mobile and AR/VR. Prophesee is based in Paris, with local offices in Grenoble, Shanghai, Tokyo and Silicon Valley.
The company is driven by a team of more than 100 visionary engineers, holds more than 50 international patents and is backed by leading international equity and corporate investors including 360 Capital Partners, European Investment Bank, iBionext, Inno-Chip, Intel Capital, Renault Group, Robert Bosch Venture Capital, Sinovation, Supernova Invest, Xiaomi.
