When Jibo, Inc., rolls out its “family robot” later this year its ability to understand and respond to spoken words will be based on the TrulyNatural(TM) embedded software from Sensory Inc. As explained in this press release, the new technology was selected by Jibo to take advantage of its ability to deliver “a natural language interface that would just work.” Jibo’s developers don’t want its users to be concerned about the quality or availability of an Internet connection to gain access to “The Cloud.”
For Sensory, it represents a reward for dozens of years of research and development of speech processing, deep neural networking and natural language understanding designed for small-footprint, low-power, embedded computer environments. Growth of the Internet of Things (IoT) promises to accelerate acceptance and licensing of its technology for the growing population of sensors and microphones in consumer electronics, mobile devices, wearables and automobiles. The company has distilled its value proposition into three words: “Security, Efficiency and Accuracy.”
In a longer form description of its innovations, the company enumerates additional advantages of the TrulyNatural. Because an internet connection is not needed, it is always available to the end user. Because processing is done “natively” and not in the cloud, personal information (including voiceprints) are secure and not shared. The large vocabulary and natural language understanding will enable individuals to take full voice control of any device or application. Local processing also removes pauses or latencies in a conversation. Plus, the company claims highly accurate recognition, even in noisy environments.
In the prepared statement, Cynthia Breazeal, Founder and Chief Scientist at Jibo, said she found Sensory’s technology and approach to be “particularly comforting” because it will deliver strong performance and accuracy without the need for an Internet connection. Indeed, a robust debate is forming among vendors, enterprise IT and security personnel and technology or security analysts surrounding the pros and cons of “on-device versus in-the-cloud” approaches.
Score one for embedded.
Categories: Intelligent Assistants, Intelligent Authentication, Articles