Syntiant: Optimizing Silicon and Deep Learning Models to Customize Voice Experiences at Edge

syntiant

Syntiant is enabling customized voice experiences at the edge, across multiple products and use cases including wake word, command control, and event detection, free from cloud connectivity, ensuring privacy and security.

Headquartered in Irvine, California, Syntiant Corp. is moving artificial intelligence (AI) from the cloud to edge devices. The company’s advanced chip solutions merge deep learning with semiconductor design to produce ultra-low-power, high performance, deep neural network processors for always-on applications in battery-powered devices, such as hearing aids, earbuds, smart speakers, mobile phones and laptops.

Backed by some of the world’s strongest strategic investors, including Intel Capital, Microsoft M12, Bosch Ventures, and the Amazon Alexa Fund, Syntiant’s deep learning “intelligent voice” solution opens new markets and business opportunities for customers as demand for speech grows globally.

The company’s highly accurate, ready-to-use, always-on voice solution adds wake word and speech control (in multiple languages) to any smart device, with ultra-low-power and greater efficiency. Other applications include sensors, event detection, and speaker recognition.

Custom built to run AI workloads, the Syntiant® NDP100™ and Syntiant® NDP101™ microwatt-power Neural Decision Processors™ (NDPs) improve efficiency by 100x and throughput by 10x vs. traditional microcontroller units (MCUs).

The company also protects user privacy and security by deploying deep learning at the edge, while lowering costs due to die size, as well as enabling implementation into the smallest devices. Syntiant also collects and trains its own data. Technical specs  on the company’s NDPs can be found by clicking on the following links:  https://www.syntiant.com/ndp100 and https://www.syntiant.com/ndp101

 

A Mission to Succeed in Disruptive Market

Founded in late 2017, Syntiant is driven by a global mission of bringing AI voice solutions to edge devices, free from touch and cloud connectivity, and that are more useful, more efficient, and more safe and secure.

The company’s deep learning voice technology is leading the way toward universal hands-free functionality, changing the way people interact with their devices.

Syntiant has moved very fast, from closing its Series A round in October 2017, to sampling its first products in June 2018, to closing its Series B round in October 2018. In September 2019, the company started shipping its Syntiant Neural Decision Processors (NDPs) to customers across the globe.

Designed into a wide range of battery-powered devices, Syntiant expects the first of these NDP-embedded consumer products to be available this year.

 

A Leadership with Futuristic Vision

Tech veteran and CEO of Syntiant, Kurt Busch is quickly building Orange County’s next great chip company. He along with Jeremy Holleman, Ph.D., Chief Scientist; Pieter Vorenkamp, COO; and Stephen Bailey, Ph.D., CTO, and Chief Architect; founded Syntiant using a clean-sheet design for AI processors to make voice the new user interface of the future.

Kurt has extensive experience in product development, having driven the successful launch of new products, ranging from SAS and semiconductors for telecom and broadcast video to consumer electronics and data center systems. Prior to Syntiant, he was President, CEO, and a member of the Board of Directors at Lantronix (NASDAQ: LTRX), a global provider of secure data access and management solutions for the Internet of Things (IoT) and information technology (IT).

Kurt has also served as a Senior Vice President and General Manager of a high-performance analog business unit at Mindspeed Technologies, acquired by MACOM. He served at the University of California at Irvine’s (UCI) alumni board for six years, dedicating his time to help graduates find success after college. Kurt is a member of the UCI CEO Roundtable and an inductee of the UCI School of Engineering Hall of Fame. He recently was recognized by UCI’s School of Engineering with the school’s Lauds & Laurels Distinguished Alumni Award.

Also, a member of the Global Semiconductor Alliance, Kurt earned Bachelor of Science degrees in electrical engineering and biological science at UCI and holds a Master of Business Administration from Santa Clara University.

 

Extending Quintessential Innovations for Industry-Demands 

Using a co-design methodology that optimizes silicon and deep learning models together, Syntiant’s architecture achieves breakthrough performance by slashing power consumption while increasing throughput. The result is an inference solution capable of intelligent processing with months of always-on life from a coin-cell battery, providing consumers a smart device experience in battery-powered edge devices.

With over 500k parameters, far more than available on typical microcontroller-based neural solutions, Syntiant’s NDP100 and NDP101 microwatt-power processors provide the neural capability for accurate inference on challenging tasks, supporting up to 64 different events that could include words, audio events such as breaking glass or gunshots, or other sensor inputs such as gas or vibration.

Syntiant’s NDPs also can serve as an always-on gate-keeper, providing a significant classification of events, allowing larger systems to remain dormant until absolutely needed. With support for multiple programmable words, Syntiant’s technology expands the voice interface from simple wake words to a command language model, enabling a local speech control interface that does not require a cloud connection, increasing security and ensuring privacy.

Syntiant’s NDPs are now available for use with the Amazon Alexa Voice Service (AVS). As a low-power processor for close-talk voice-enabled devices, the company’s neural network technology allows customers to offer a built-in Alexa experience to even the most power-sensitive battery-powered devices.

 

Moving Towards an AI-Driven Reality

Kurt believes that “As our world becomes increasingly interconnected through data and technology, we are inching closer to a reality that’s driven by artificial intelligence, the merging of our physical and digital lives.”

According to him, Edge AI is at the heart of this paradigm shift. More importantly, Syntiant’s technology is enabling customers to bring deep learning to battery-powered devices at the edge with significantly less power consumption than current MCUs and without cloud connectivity.

In remote parts of the world, for example, Syntiant’s NDPs easily can enable the lowest-cost feature phone to listen to the human voice, enable a voice assistant and use the Internet just like a very high-end smartphone.

Similarly, high-performance neural networks at the edge are essential for demanding public safety environments and industry operations. According to Scott Mottonen, Corporate Vice President, devices at Motorola Solutions, “Syntiant’s deep learning neural decision processors can provide users working remotely in the field with enhanced safety and efficiency by powering a new breed of applications at the edge.”

The company’s NDPs incorporates an optimized neural network architecture supporting more than 500,000 weights to bring significant neural processing power to a very small energy footprint. This level of neural processing enables voice interfaces in even the harshest environments.

At the nexus of technology and social reform, lies the concern over security and privacy. By running inference on the local device, the company’s neural technology avoids the privacy risk incurred by transmitting personal data to potentially insecure servers.

From enhanced functionality and better productivity to improving safety and privacy, Syntiant’s technology is impacting innovation, making AI at the edge a reality.

 

Advancing with New-Age Technologies and Significant Partnerships

Deep learning has provided a toolset that yields vastly superior decisions to traditional methods based on human intuition, essentially making devices smarter through the use of AI voice solutions.

However, deep learning methods are computationally intensive. Traditional semiconductor solutions for advanced deep learning at the edge are severely limited by excessive power dissipation.

With a power consumption of less than 140µW and package sizes down to 1.4mm X 1.8mm, the company’s neural decision processors provide battery-powered consumer products typically 100x more efficiency and 10x the throughput of the typical low-power MCU solution.

Here are some of the significant partnerships:

Amazon Alexa Voice Service (AVS): Syntiant® NDP100™ and Syntiant NDP101™ microwatt-power Neural Decision Processors™ (NDPs) are available for use with the Amazon Alexa Voice Service).

Qualcomm: Syntiant is a member of the Qualcomm® Extension Program, offering OEMs and ODMs accelerated integration of its ultra-low-power neural network technology with Qualcomm’s audio platforms and systems on chip (SOCs).

Sensory Inc. – Syntiant has partnered with Sensory Inc., AI, and speech recognition pioneer, to deliver a deep learning, multi-lingual voice interface for battery-powered devices. The combined solution merges Sensory’s larger, higher-performance TrulyHandsfree™ wake word engine and voice control with Syntiant’s microwatt-power Neural Decision Processors™ (NDPs), bringing low-latency, real-time inference to edge devices, enabling consumers to seamlessly “voice” commands in dozens of languages.

Infineon Technologies: Syntiant also has partnered with Infineon Technologies, a world leader in semiconductor solutions, to service the rising demand for the voice user interface (VUI) in electronic devices. The joint solution enables developers to easily create custom, always-on voice user interfaces, allowing exact voice pick-up from long-distance, with whispered speech or in noisy environments.

MSI: Syntiant has partnered with MSI, the global leading gaming brand, to develop edge AI computing, giving users access to voice command features regardless of cloud connectivity.

 

Prestigious Awards and Testimonials

Syntiant is the first and only commercial silicon designed to put deep learning processing into almost any consumer device.

Syntiant’s award-winning technology was developed by top minds in semiconductors and deep learning, receiving numerous accolades, including: Fast Company’s prestigious list of the World’s Most Innovative Companies for 2020; The Linley Group’s 2019 Analysts’ Choice Award in the Best Embedded Processor category; CES® 2020 Best of Innovation Awards Honoree; “Outstanding Emerging Technology Company” by OCTANe; Best Deep Learning Company by AI Breakthrough; CB Insights AI 100 List of Most Innovative Startups; and 451 Research’s “451 Firestarter” Award; among others.

Here’s what others have said about Syntiant:

“Syntiant’s neural decision processor technology takes a unique approach to wake word detection on voice-enabled devices,” said Sanjay Voleti, Senior Manager of device enablement with the Alexa Voice Service. “We recently conducted acoustic performance tests of their latest solutions and were impressed by their ability to support compute-on-edge and wake word detection on low-power devices. We’re excited to see developers begin using this technology in their devices and deliver new Alexa experiences for customers.”

“As a member of our Qualcomm Extension Program, Syntiant can combine its always-on word recognition and event detection solution with our ultra-low-power Bluetooth audio platforms,” said Chris Havell, Senior Director, Product Marketing, Qualcomm Technologies International, Ltd. “Technologies like this enables powerful compute-on-edge processing to be offered in headsets, earbuds, and hearables without an Internet connection and with very low power consumption.”

“Infineon´s best-in-class XENSIV™ MEMS microphones are redefining the acoustic, electric power consumption and robustness standards for voice user interfaces,” said Dr. Roland Helm, Ph.D., Head of Product Line Sensors at Infineon Technologies. “Together with Syntiant we are facilitating clear voice commands at the edge that support audio and keyword classifications, on and off-line, in a variety of challenging, power-critical always-on and battery-driven environments.”

 

A Progressive Roadmap to Prosperity

According to Kurt, Syntiant is creating real value with its deep learning technology, making edge devices smarter with its AI voice solution, as global consumer demand shifts from touch screens to speech control.

The company believes consumers soon will be controlling all kinds of edge devices with its built-in voice technology.

Given the ease-of-use of its solution for putting deep learning into battery-powered devices that is a low cost, cloud-free, low power, and super-efficient, Syntiant is uniquely positioned to address the growing AI voice market opportunity.