AI News, Eta Compute Debuts Spiking Neural Network Chip for Edge AI
Eta Compute Debuts Spiking Neural Network Chip for Edge AI
At Arm Tech Con today, West Lake Village, Calif.-based startup Eta Compute showed off what it believes is the first commercial low-power AI chip capable of learning on its own using a type of machine learning called spiking neural networks.
A neural network that can do what’s called unsupervised learning can essentially train itself: Show it a pack of cards and it will figure out how to sort the threes from the fours from the fives.
“When you talk about the democratization of machine learning and getting regular engineers and computer scientists involved in IoT, you need to be able to make the training [of networks] simpler.” Because their structure includes feedback, spiking neural networks are capable of training themselves even without a labeled set of data.
The TENSAI chip consists primarily of an Arm M3 core and an NXP CoolFlux digital signal processor core.The DSPhas two arrays dedicated to doingdeep learning’s main computation—multiply and accumulate.These cores are implemented in a manner that’s called asynchronous, subthreshold technology.
“We spenta lot of engineering time on analog;it’s a source of advantage.” The result is a chip thatEta Compute believes will be miserly enough to listen outfor “wake words” in battery operated audio devices.
BrainChip Enters AI Territory with Spiking Neural Network
It may take a machine to learn about all of the new machine-learning techniques popping up these days.
BrainChip is adopting a neuromorphic computing approach that’s different from the other deep neural network (DNN)/convolutional neural network (CNN) solutions already available.
CNNs use a high-overhead feature-learning process—it generates weights used in a classification process that can identify multiple items.
The threshold logic and connection reinforcement of SNNs both exhibit high efficiency, allowing even software implementations to work on lower-end hardware.
Convolutional neural networks tend to be math-intensive with more complex training requirements, while spiking neural networks have less overhead but target more specific classification results.
The SSN approach delivers a 7X improvement compared to GPU-accelerated deep-learning classification neural networks like GoogleNet and AlexNet in terms of frames/s per watt.
For example, a particular backpack or person may need to be tracked through video streams from dozens of cameras for security purposes.
The initial image resolution can be as low as 20 by 20 pixels for patterns and 24 by 24 pixels for facial recognition.
BrainChip Takes Spiking Neural Networks to the Next Level
Neural networks rank as the hottest machine-learning (ML) trends in artificial-intelligence work these days.
CNNs and other deep neural networks (DNN) usually involve a training process that generates a model that can be used for inference.
Leveraging feed-forward training, spiking neural networks have low computational and power requirements compared to CNNs.
CNNs and SNNs are similar in that they take input, analyze the input using their neural-network model, and generate responses.
Developers convert input data into a form that’s usable by the algorithms, and that often means preprocessing the raw data.
BrainChip provides a number of data-to-spike converters for developers to start with, but more can be created depending on the application and source data format.
SNNs emit spikes in a somewhat similar fashion, but spikes aren’t always generated at each point, depending on the data.
SNNs developed with Akida will be able to run on the hardware when it becomes available, although some applications may work well without resorting to specialized hardware.
The development tools are integrated with the NumPy Python numerical package for massaging input and output data as well as matplotlib, a Python 2D plotting library.
- On Sunday, January 20, 2019
The Future of Deep Learning Research
Back-propagation is fundamental to deep learning. Hinton (the inventor) recently said we should "throw it all away and start over". What should we do?
ISSCC2018 - Brain-Inspired Technologies: Towards Chips that Think
Barbara De Salvo, Deputy Director for Science and Long Term Research, CEA-Leti, Grenoble, France Since the late 50s, brain-inspired computing has been ...
September Event 2018 — Apple
Apple Special Event. September 12, 2018. Big news all around. Take a look at the all-new Apple Watch Series 4, iPhone XS and iPhone XS Max, and iPhone ...
A look at methods and concerns for augmenting the human mind. Use my link and get a free audio book with a 30 day trial
How Can the Brain Efficiently Build an Understanding of the Natural World
CBMM Special Seminar Ann M. Hermundstad, PhD, Janelia Research Campus Abstract: The brain exploits the statistical regularities of the natural world. In the ...
Biology Is Technology | Oshiorenoya Agabi | TEDxVilnius
Alive computers and living neurons: in his talk Oshiorenoya shares insights on the future of neuro-computing: next generation computational platforms with ...
What Is Cognitive Computing (How AI Will Think)
This video is the eleventh in a multi-part series discussing computing. In this video, we'll be discussing what cognitive computing is and the impact it will have on ...
Dockside Diplomacy | Critical Role | Campaign 2, Episode 35
Thanks to D&D Beyond for sponsoring this episode of Critical Role! The Mighty Nein find success in their task for the Ruby, but chaos ensues as they continue ...
The Echo Tree | Critical Role RPG Show Episode 63
New upload to correct corrupted version *** Check out our store for official Critical Role merch: Catch Critical Role live Thursdays at 7PM ..
Alex Guarnaschelli: "Old-School Comfort Food: The Way I Learned to Cook" | Talks At Google
Alex Guarnaschelli grew up in a home suffused with a love of cooking, where soufflés and cheeseburgers were equally revered. The daughter of a respected ...