Sensory launches platform for speech recognition at network edge to protect customer privacy

Sensory has announced the first full-feature release of its embedded large vocabulary speech recognition platform, TrulyNatural, to help companies create products with natural language interaction without compromising customer privacy or taking on the high memory cost of open-source based technologies.

TrulyNatural is the product of extensive development and five years of beta testing, according to the announcement. A survey published by PCMag.com in March indicated that privacy is the top concern related to smart home devices for the highest number of consumers, and that was before controversy broke out with reports that Amazon employees listen to Alexa recordings as part of the company’s algorithm training process. TrulyNatural addresses privacy concerns by processing voice data on the device, and never storing voice requests.

“To benefit from the advantages afforded by cloud-based natural language processing, companies have been forced to risk customer privacy by allowing always listening devices to share voice data with the recognition service providers,” comments Sensory CEO Todd Mozer. “TrulyNatural does not require any data to leave the device and eliminates the privacy risks associated with sending voice data to the cloud, and as an added benefit it allows product manufacturers to own the customer relationship and experience.”

Sensory’s technology is an order of magnitude smaller than open source options, but does not sacrifice performance, with a task completion rate higher than 90 percent in internal testing, the company says. Sensory acquired benchmarking and accuracy assessment company Vocalize.ai earlier this year.

In a natural language setting with a large vocabulary, TrulyNatural can be as small as 10MB, while for applications like virtual assistants and call center chatbots it can be scaled to support a larger vocabulary. The categorization method of TrulyNatural also does not require scripted grammars. TrulyNatural can be deployed for home appliances, vehicle infotainment systems, set top boxes,

Because TrulyNatural processes speech at the edge, rather than in the cloud, it does not require an internet connection. Product manufacturers can also combine TrulyNatural with Sensory products like TrulyHandsfree wake words for branded or custom user experiences, or TrulySecure for voice biometric access control.

“Consumer concerns over security and privacy have been growing over time and Sensory’s TrulyNatural platform addresses this by embedding natural language speech recognition locally on device,” says Dennis Goldenson, Research Director, Artificial Intelligence and Machine Learning with SAR Insight and Consulting. “As a result, TrulyNatural improves response time and delivers a high performing, more secure and reliable solution. Product manufacturers will appreciate TrulyNatural’s speech engine technology because it enables them to implement a highly valued voice experience through their own brand name and avoid surrendering customers to a potential competitor.”

TrulyHandsfree TrulyNatural supports U.S. English, with support for UK English and 10 other languages planned for release in 2019 and 2020. SDKs are available for leading platform including Android, iOS, Windows, and Linux.

Moving voice recognition to the edge is the motivation for a recent partnership between Snips and DSP Group, and a new line of AI SoC sensors from AIStorm.

Article Topics

biometrics  |  biometrics at the edge  |  privacy  |  Sensory  |  speech recognition