DF Robot is shipping a CurieNeurons Kit , a toolbox for makers who want to have a shot at adding AI to their IoT projects. The kit includes an Arduino/Genuino101 board and a collection of sensors and actuators. Using the General Vision CurieNeurons Pro library, Arduino developers can train the neurons inside the Curie chipset by showing examples and query them to classify signals without worrying about the technical nuances of how neural networks actually work.
nepes, leader in advanced semiconductor assembly and packaging technologies, and General Vision (GV), leader in the design of digital neural network and low-power AI components, have signed an agreement to develop and manufacture a new wafer chip-scale packaged (WCSP) neuromorphic chip, NM500 with 576 neurons. GV will design the chip based on its NeuroMem® technology, and nepes will have the exclusive rights to manufacture, package and sells this chip worldwide.
The NM500 will have the unique NeuroMem architecture, functionalities and properties including trainability on-the-chip, fixed latency, parallel interconnectivity and low power. It can learn and recognize patterns deriving from text, scientific datasets, bio-signals, audio files, images, and videos, etc. Networks of different capacity can be easily assembled by connecting multiple Chip-Scale Packaged NM500 in parallel. Nepes package design capabilities will allow to manufacture a variety of System-On-Chips (SoCs), System-in-Packages (SiPs) and modules, whether combining neurons and processor in a same chip, or assembling large capacity of NeuroMem networks in a same package.
The 1st samples of NM500 are expected for Q2 2017 with mass production starting Q3 2017. In addition, Nepes and General Vision will keep working together developing products and expanding business areas for neuromorphic products and applications. The NM500 will be the third commercial chip featuring NeuroMem neurons after GV’s own CM1K chip (1024 neurons) and the pattern recognition accelerator of the Intel® Curie/QuarkSE module (128 neurons).
Read the complete press release.
JD Yoon, Jung-Ho Ahn, Anne Menendez, Guy Paillet, Jeff Woo
Mando-Hella Electronics and General Vision announce a collaboration to develop an Advanced Driver Assistance and Monitoring system powered by NeuroMem neural networks to deploy fast, adaptive and low-power multi-sensor recognition engines. The project will cover multiple aspects of the automated driver assistance including driver awareness monitoring and obstacle detection.
Mando-Hella Electronics (MHE) is the first Korean tier1 supplier which manufactures sophisticated and safe automotive systems such as ABS, ESP, ECS and EPS.
General Vision has signed a distribution agreement with Maker Collider, leading maker community in China to deliver a suite of libraries and tools unleashing the neurons of the Intel Curie module on the Arduino/Genuino 101. The libraries open up the access to the 128 neurons inside the Curie module which are based on the NeuroMem technology pioneered by General Vision. They will allow developers to train everyday objects to perform interesting tasks such as teaching the children how to brush efficiently their young teeth, monitoring health of dogs or favorite animals and much more to come.
“Artificial Intelligence should be an innovation tool in everyone’s hands, you cannot imagine this before the debut of Curie and CurieNeurons, we will work with Intel and General Vision to make it happen, to let people design their intelligent devices using not coding but training.” says Honggang Li co-founder and CEO of Maker Collider.
“General Vision is delighted to team with Maker Collider and Intel to allow the development knowledge in new domains via training instead of programming.
General Vision is releasing a NeuroMem API for the Intel Arduino/Genuino101 board. This embedded module features a QuarkSE microcontroller with a pattern recognition engine composed of 128 NeuroMem neurons. Using the new API, developers can teach and query the neurons in minutes to monitor and classify signals received from the on-board MEM or other sensors. The API is delivered with examples and a getting started guide.