BrainChip offer robust neural network converter

• Convert existing Convolutional Neural Networks to high performance, low power event-based Spiking Neural Networks
• Innovation in Neural Network design enables the next generation of AI Edge devices

BrainChip Holdings Ltd, leading AI Edge has debuted its powerful neural network converter. This device enables users to easily convert existing convolutional neural networks (CNNs) to a high-performance, low-power event-based spiking neural network (SNN).

BrainChip’s SNN is integrated with the Akida Development Environment to provide network conversion and simulation.

“The low power inherent in the Akida device will set a new standard in neural network design, implementation and performance,” said BrainChip CEO, Roger Levinson.

The device is the world’s first commercially available development which enables CNN and SNN implementations on the same hardware device without comprising performance and power benefits of networks.

Users can achieve higher performance with a native SNN and faster time-to-market by using the converter and the ADE.

The new conversion flow is meant to be easy and convenient to use, whilst maintaining accuracy. Users can implement image processing anomaly detection, ADAS, robotics and keyword spotting.

The conversion flow takes an input as a standard CNN description which the user can modify to Akida compatible layers. The modified network description is then processed through standard training.

“Edge solution developers 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,” Rodger said.

The Akida development environment, inclusive of the CNN to SNN conversion capability is available today in beta release. General release is expected in Q3 2019.