BrainChip: Lean machine learning for edge network devices

An Intellyx Brain Candy Brief

When I heard about a ‘neuromorphic, spiking neural network edge’ device vendor called BrainChip, I thought, the silicon/brain interface was finally here and we could tap it into our skulls and jack into the matrix. Boy, was I wrong!

What BrainChip does offer is the Akida Neuromorphic System-on-Chip, a small footprint chip with its own embedded CPU and memory. By using a technique known as ‘neural spiking,’ they claim that these chips can autonomously learn to recognize trends and patterns faster than other deep learning approaches, and build AI models with far less power consumption.

“Most deep learning systems push a lot of data pollution back, but neural spiking learns to make inferences on the edge that remove irrelevant data from the model,” said Louis DiNardo, CEO, Brainchip.

The Akida embedded ML chipset and developer toolkit might have utility for any scenario involving decisionmaking or alerting for IoT devices or edge computing scenarios, from recognizing behaviors in smart surveillance footage, to autonomous drone and vehicle movement, human sight and sound augmentation, responsive medical device applications, and more.

Just please, don’t stick it in your brain.

© 2019 Intellyx. At the time of writing, BrainChip is not an Intellyx customer. Want to see more BrainCandy? Subscribe today. If you are a vendor seeking coverage from Intellyx, please contact us at PR@intellyx.com.

SHARE THIS:

Principal Analyst & CMO, Intellyx. Twitter: @bluefug