With 2016 fast approaching, expectations are high for the Internet of Things (IoT). Just don’t ask me to build a consensus around the meaning of the term. My favorite, recent description of the IoT went something like this: “Whether it’s a TV, lightbulb, automobile, refrigerator or robot – if it is powered up, it’s on the Internet”. There’s an element of hype to this grandiose description, but it also reflects hope, as well as the need for simple ways for smartphones, PCs, laptops, tablets and even automobile dashboards to help us take control of digital devices and applications

What’s lacking is a natural user interface (NUI) for the IoT.

“Things fall apart” is a meme that captures the modern world’s entropic tendency toward the chaotic. The vision of an IoT transforms the Internet into an interstitial fabric between people and all of their connected possessions. Each “thing” in the IoT will need to respond to human instructions if it is going to be a true personal assistant. It feels like a daunting task to build the natural interfaces that will enable humans to take command by using their own words.

Cue Nuance Mix. Formally launched today, it is Web-based development environment that enables software and app developers to add speech recognition, natural language understanding (NLU) and conversational interactions to help individuals take command of the IoT. It enables basic Web and mobile app developers to do so without requiring a PhD in computational linguistics or artificial intelligence. As demonstrated in the video below, it learns to understand concepts, give responses or take actions based a a very limited amount of input.

In the video, Nuance provides a use case that involves a home robot. It has a number of capabilities, but cannot respond to commands because it doesn’t understand spoken or keyed-in commands. The demo shows how the robot “learns” what people are telling it to do through natural language processing. It doesn’t take much for it to master a concept and, once that is done, it can keep refining its ability to understand and take action.

Nuance mix tackles a well-known problem head-on. Much of what the IoT promises to do is complex. Turning down a thermostat or turning off a light can be handled by basic rules supported by platforms like IFTT (If This Then That), which puts the burden on individual end users to discover and enter their own rules to live by. More complex tasks, like finding a cat or even telling a home entertainment system to find and queue up English Premier League soccer (or football) for some binge watching, can be much more complicated, especially if you want to instruct the system with spoken commands.

Nuance Mix is a development environment that integrates a stack of speech processing, natural language understanding, and machine learning resources and uses it to provided a way for developers and (end-users) to train it with their specific domain knowledge. That’s the magic of this model. It’s open brings proven speech processing technologies into the mix (so to speak). The Internet is not the great equalizer that it is made out to be. Natural Language Understanding may turn out to be. Nuance Mix brings NLU to the broad application developer community. It is available as a limited release beta through the Nuance Developers program. It is a full stack of Nuance speech processing and natural language resources.

Categories: Intelligent Assistants Conference, Articles

Tags: speech processing, Artificial Intelligence, Natural language processing, Internet of Things, intelligent Assistants