NEWSBYTE: ARM launches scalable chips for IoT machine learning

Semiconductor and software giant ARM has announced a new range of scalable processors designed to deliver enhanced machine-learning capabilities to IoT devices.

The Softbank-owned company, which counts Apple, Samsung, drone maker DJI, and FitBit among those who rely on its architecture, is pitching Project Trillium as the industry’s most scalable, versatile ML compute platform.

The Project Trillium suite will, according to Rene Haas, ARM’s president of IP Products Group, aim to find a better balance between energy efficiency and computing power.

“The rapid acceleration of artificial intelligence into edge devices is placing increased requirements for innovation to address compute, while maintaining a power efficient footprint,” he said.

“To meet this demand, ARM is announcing its new ML platform, Project Trillium. New devices will require the high-performance ML and AI capabilities these new processors deliver. Combined with the high degree of flexibility and scalability that our platform provides, our partners can push boundaries of what will be possible across a broad range of devices.”

• SoftBank also owns Aldebaran Robotics, makers of the NAO, Pepper, and Romeo humanoids, and robotic giant Boston Dynamics, formerly part of Alphabet.

Internet of Business says

The edge environment is an increasingly important part of the IoT, and with AI and machine learning being embedded into more and more functions that need to execute in real time, these type of innovation can only grow.

The post NEWSBYTE: ARM launches scalable chips for IoT machine learning appeared first on Internet of Business.

Internet of Business

MIT develops low-power high-speed chips for IoT security

MIT chip performs hardwired encryption faster and using less power.

MIT researchers have hardwired public-key encryption into a new chip for IoT devices. It uses 1/400 of the power of software execution, one tenth of the memory, and executes 500 times faster. 

From data breaches to weaponised devices, the Internet of Things (IoT) has been plagued with security issues. In part, this is down to hardware manufacturers implementing security as an afterthought, along with a lack of standardisation.

But it’s also true that building a low-power network of connected devices will remain challenging while encryption is so energy intensive.

Sensitive data transactions are usually protected by public-key cryptography. This type of encryption allows computers to transfer information securely without needing to establish a secret encryption key.

However, the software responsible for executing these protocols is both memory and energy intensive. The battery life trade-off required for embedded sensors and smart devices to run has long been a burden on development.

Read more: Virtuosys launches Edge Application Platform

Energy-efficient encryption for the IoT

But that could be about to change. Researchers from MIT have developed a chip that’s hardwired to execute public-key encryption.

It uses a tiny fraction of power (1/400) compared with software execution of the same protocols, and just ten percent of the memory. Better still, MIT’s new chip executes the encryption process 500 times faster.

The new chip relies on a technique called elliptic-curve encryption. The process uses mathematical functions to secure transactions. Previously, chips have been hardwired to handle specific elliptic curves or families of curves. MIT’s latest chip has been developed to work with any elliptic curve.

“Cryptographers are coming up with curves with different properties, and they use different primes,” said Utsav Banerjee, an MIT graduate student in electrical engineering and computer science, and lead author on the paper.

“There is a lot of debate regarding which curve is secure and which curve to use, and there are multiple governments with different standards coming up that talk about different curves. With this chip, we can support all of them, and hopefully, when new curves come along in the future, we can support them as well.”

The researchers will present a paper on the new chip at this week’s International Solid-State Circuits Conference.

Internet of Business says

This is merely the latest innovation from MIT to focus on reducing the energy consumption of intelligent systems, while increasing their power and speed. Our separate report today looks at its work with neural networks. Energy use, cost, and speed are the critical elements in developing sustainable IoT devices and services that can really deliver on their promise.

Read more: NEWSBYTE: ARM launches scalable chips for IoT machine learning

Read more: Dell Technologies unveils new IoT strategy in New York

Read more: MIT’s NanoMap helps drones to navigate safely at high speed

The post MIT develops low-power high-speed chips for IoT security appeared first on Internet of Business.

Internet of Business

Apple buying Chinese storage chips for iPhones ‘in Chinese market only’ – Nikkei

A brief report claims that Apple will be adding a Chinese NAND storage chip supplier to its roster. The company has in the past bought these chips from Toshiba, SK Hynix and Samsung, but is now said to be in talks with Yangtze Memory Technologies.

The report is light on detail, but does make one interesting claim …



Fast talker: Alexa may offer speedier answers with Amazon-made AI chips

Amazon wants to cut the lag time between your asking Alexa a question and the virtual assistant giving you an answer. According to a report by The Information, the online retailer is developing its own artificial intelligence chips to be used in Echo devices and other hardware. If successfully created and deployed, these AI chips would allow more voice-based requests to be processed on-device rather than going to the cloud.

Currently, Alexa needs to contact the cloud to interpret commands. That’s why there’s a short delay after you ask the virtual assistant a question—it needs to analyze the command and gather an answer with help from the cloud. A dedicated AI chip in a device like an Echo would allow Alexa to process certain requests more quickly, decreasing the delay that lies in between your question and Alexa’s answer. While complex inquiries will likely still be handled with help from the cloud, more simple commands could be processed all on the device itself.

Amazon reportedly has 450 people with chip knowledge on staff now, many of which came via recent acquisitions. The company bought the Israeli chipmaker Annapurna Labs in 2015 for $ 350 million and the security camera company Blink for a reported $ 90 million at the end of 2017. It’s believed that Amazon bought Blink specifically for its low-energy chip expertise; the company’s smart home security cameras use these chips to extend the battery life of its camera modules to at least two years.

Read 2 remaining paragraphs | Comments

apple – Ars Technica

Amazon may be developing AI chips for Alexa

 The Information has a report this morning that Amazon is working on building AI chips for the Echo, which would allow Alexa to more quickly parse information and get those answers. Getting those answers much more quickly to the user, even by a few seconds, might seem like a move that’s not wildly important. But for Amazon, a company that relies on capturing a user’s interest in… Read More
Mobile – TechCrunch