By Charles King, Pund-IT, Inc. November 6, 2019
Internet of Things (IoT) edge computing is often discussed as a single entity, but it is more diverse and complex than many might assume. A sizable portion of edge computing offerings leverage gateway systems that collect, coalesce and analyze sensor data before sending it on for further analysis and decision making. But increasingly sophisticated processor and software technologies, and artificial intelligence (AI) functions and features are transforming edge computing in both business and consumer endpoint devices.
A good example of this is the recent partnership between AI accelerator chip and technology provider Gyrfalcon Technology Inc. (GTI) and AI solution provider Sensory to deliver low latency/high reliability embedded machine learning capabilities on edge devices. Let’s take a look at what GTI and Sensory provides individually, and what they’re planning to deliver via their partnership.
GTI and Sensory alone
GTI developed its first Edge AI chip in 2017 and has long focused on enabling high performance, low cost, energy efficient AI acceleration for mass market devices. Since its launch in 2017, the company has brought to production neural network acceleration technologies, including the Lightspeeur 2803S AI processor that delivers 24 TOPS/W for high performance at low power. Vendors have been developing mobile and edge AI devices that utilize GTI’s AI Accelerator solutions. In fact, last month LG announced that it will incorporate GTI chips in its new Q70 Smartphone to enable affordable premium image and video features.
Sensory has over two decades of experience developing biometric technologies, including machine learning technologies that support voice, vision and natural language processing. The company’s embedded speech recognition solution (including a well-engineered and patented codebase) has shipped in over 2 billion consumer products. Additionally, Sensory recently perfected an algorithm that can recognize legitimate user faces with the highest levels of accuracy and is central to the company’s TrulySecure authentication solution.
That new algorithm can fully leverage the low latency and high reliability features in GTI’s AI Accelerator chip portfolio, enabling devices to recognize legitimate user faces quickly and accurately. Additionally, the companies’ technologies are particularly well-suited to one another. According to Sensory’s CEO Todd Mozer, “Our TrulySecure solution can run up to four times faster and provide a larger, more accurate model thanks to GTI’s AI accelerator chip.” As a result, endpoint devices such as smart phones that feature the GTI/Sensory offering should support safer and more secure transactions, and also enhance other authentication-focused solutions and services.
The far edges of the Internet of Things are more complex and diverse than many might assume. That’s partly because IoT edge environments continue to evolve as businesses and consumers consider the challenges, needs and use cases that IoT edge solutions can address. In turn, the potential of IoT edge offerings is being improved by the efforts of innovative vendors, like Gyrfalcon and Sensory.
The two companies have both achieved exceptional results individually but by working together, they appear ready to significantly enhance the quality, speed and security of authentication functions featured on smart phones and other mobile and edge computing devices. Those achievements should obviously benefit GTI and Sensory. But they will also measurably and positively impact the organizations and end users that depend on edge computing and IoT solutions.
© 2019 Pund-IT®. All rights reserved.