How does Mix.nlu work?
Imagine that you’re a developer who’d like to speech enable your smart home device. To build a really great experience, you’ll need the speech interface to be fully customized to handle all the skills that make your device unique.
Vincent Vac chase the white cat. This is something that a user might say to your device.
This is what you will make happen.
Mix.nlu understands what the user said and provides you a set of user intents and concepts in {json}.
How you make 
Mix.nlu Works For You Now that you know what you can do with Mix.nlu, let’s talk about how you’ll do it.
This is the structure you will set up in Mix You define your own intents and concepts with Mix.nlu  intent
 Sample
Chase
1. Define your intents. An intent represents an action or domain available within your device. You can think of intents as the actions Mix.nlu will return to your app via json, which you can map to your existing functions.
Vincent Vac to chase the white cat 2. Teach Mix.nlu how your users might express each intent. Samples are sentences or phrases a user might say. At first, you have to teach Mix.nlu how to classify and map sample sentences to intents. With enough samples, Mix.nlu will start to recognize language patterns so it can automatically understand new samples it’s never seen before. This is the power of natural language understanding. Vincent Vac to go after the cat Vincent Vac chase the purple llama Hello Vincent, please chase that cat Congratulations! You now have a simple NLU Model.
A concept helps you capture additional details relevant to an intent, and how that intent should be handled For example, you could define a concept for items you’d like Vincent Vac to chase, with literals-value pairs defined for each item. YOU CAN GO FURTHER... 3 (optional). Consider other details you want to capture. Concepts
(Chasee, Color ){
Example of a Concept Concept: Chasee Literal(s) Value
cat, cats CAT fish, goldfish FSH dog, puppy DOG llama LLM turtle TTL
Tagging Concepts You can then associate or “tag” the concepts that show up in your sample sentences to the appropriate concept names you’ve already defined. Once you’ve done this a few times, Mix.nlu learns that the values within a concept are interchangeable, and it can even learn new values automatically.
Tell Vincent vac to chase the [color] white[/] [chasee]cat[/]

 

Now it’s your turn. Join us in the beta program and build something amazing. Mix.nlu makes it easy to build your own custom speech interface in a matter of minutes. And with the power of natural language, Mix learns quickly from your samples and can understand many different command variations users might speak.

Nuance Developers Home    |    Legal Notice    |    Privacy Policy    |    Policies

© 2016 Nuance Communications, Inc. All rights reserved.