TrulyNatural™ is the first embedded large vocabulary continuous speech recognizer system which can provide the same state-of-the-art performance as cloud based technologies, but on the device. Many applications don’t need, or don’t want to rely on, a cloud based connection to do high performance speech recognition. Connections can frequently fail or cause fatal delays, and recent concerns have been expressed about sending personal data to the clouds where it could be stolen or used for purposes undesirable to the owner.
Sensory uses a unique form of a neural network with deep learning to achieve acoustic models an order of magnitude smaller than the present state-of-the-art. These neural networks employ the most recent breakthroughs in speech feature extraction to produce superior accuracy in real world noise. Combining this with a small Finite State Transducer (FST) enables storage and accurate processing of multiple large search domains as required in today’s market.
New additions to TrulyNatural speech recognition allows animated chatbots with no cloud connection
Sensory, the leader in on device AI technology is now bringing chatbots to consumer electronics through new additions to its TrulyNatural line of embedded natural language speech recognition engines. Sensory’s embedded artificial intelligence line-up includes a variety of neural net deep learning approaches such as wake word technology for Alexa and other leading cloud engines, natural language voice interfaces, biometrics, and computer vision.
Sensory’s embedded AI technology lineup includes:
- TrulyHandsfree – wake words and small vocabulary low power solutions
- TrulySecure – biometric authentication with face and/or voice
- TrulyNatural – an embedded natural language large vocabulary speech recognizer
TrulyNatural has been augmented to include both a chatbot engine which can be used for dialog management and scripting, and a new talking avatar technology to enable accurate avatar movements including lip-synchronization. TrulyNatural can now enable consumer products and applications to have a voice driven interface that offers a more conversational style and the technology can further take advantage of embedded text-to-speech to drive an avatar interface.
Sensory’s avatar interface utilizes a patented non-linear morphing technology to allow facial and mouth movements between Visemes (visual representations of phonemes) to look extremely natural and realistic, even though they are completely automated.
“Sensory is always working towards the future of consumer interfaces by utilizing AI techniques that can reside on device and run on local processors for speed, portability, and flexibility. We now have the ability to put together a plethora of technologies including text-to-speech, natural language understanding, chatbot, and avatars to create a totally new kind of visual voice interface on consumer products and apps.” said Todd Mozer, founder and CEO of Sensory.
About Sensory Inc.
Sensory Inc. creates a safer and superior UX through vision and voice technologies. Sensory’s technologies are widely deployed in consumer electronics applications including mobile phones, automotive, wearables, toys, IoT and various home electronics. To date, Sensory’s technologies have shipped in over 2 billion units of consumer products from leading brands.