Sensory TrulyNatural Enables Natural Language Understanding at the Edge
May 02, 2019
TrulyNatural?s intelligent natural language user interface reduces privacy and security concerns as voice requests are never stored and never leave a local device.
SANTA CLARA, CA. Sensory Inc. has released the first full-featured version of TrulyNatural, a large-vocabulary, on-device speech recognition engine capable of natural language understanding. TrulyNatural’s intelligent natural language user interface reduces privacy and security concerns as voice requests are never stored and never leave a local device.
TrulyNatural categorizes speech into unlimited intents and entities, allowing the natural language understanding component to interpret any speech natively without the need for scripted grammar. The software is an order of magnitude smaller than open source speech recognition alternatives with possible footprints of less than 10 MB. Meanwhile, the technology delivers similar or better performance, with a task completion rate that provides greater than 90 percent accuracy.
The product can run on a single applications processor and does not require an internet connection, but can also be scaled to broad-domain applications with virtually unlimited vocabulary.
TrulyNatural currently supports US English, with UK English, French, German, Italian, Japanese, Korean, Mandarin Chinese, Portuguese, Russian and Spanish. SDK's are available for Android, iOS, Windows, Linux and other leading platforms.