In a beta release, US-based artificial intelligence device company BrainChip Holdings (ASX: BRN) has debuted its powerful neural network converter, with the general release scheduled for the September quarter of this year.
BrainChip develops advanced AI-based computing solutions that are inspired by the biology of the human neuron.
The company’s neural network technology can learn autonomously, evolve and associate information – just like the human brain.
Additionally, intelligence and analytics occur at the point of acquiring data rather than being transferred to a central processing core in a data centre or the cloud for analysis.
The company has developed the Akida chip, which is scalable and allows users to connect many Akida devices together to carry out complex network training and inference for multiple applications.
BrainChip’s new neural network converter enables users to convert existing convolutional neural networks (CNN) to an Akida compatible event-based spiking neural network (SNN).
The company has integrated the converter with the Akida Development Environment (ADE), which is a complete machine learning framework for the Akida chip. The integration of the converter to provides network conversion and simulation.
“The low power inherent in the Akida device will set a new standard in neural network design, implementation and performance,” BrainChip chief operating officer Roger Levinson said.
“With no barriers to utilising the Akida platform, edge solution developers will have their cake and eat it too.”
“They can leverage existing CNN solutions as well as incorporate next generation SNN solutions all in a single development environment and on a single device and achieve a low power solution without sacrificing performance,” Mr Levinson explained.
Akida technology applications
According to BrainChip, the company’s technology is suitable for numerous edge applications including surveillance, advanced driver assistance systems, vision guided robotics, drones, internet of things, acoustic analysis and cybersecurity.
The Akida chip includes BrainChip’s entire AI edge network and has multiple learning modes. Additionally, the new CNN to SNN conversion flow has been designed for ease of use and employs standard text files.
“The conversion maintains near full accuracy with increased performance while eliminating neural network computational overhead,” the company noted.
“Post-conversion, the entire network is executed within the neural fabric of the Akida chip, which means that the host computational requirements of the neural network are eliminated.”
Today’s news pushed BrainChip’s share price up almost 23% to reach $0.075 by mid-morning trade.