botframework Adding Natural Language Processing Defining a LUIS Model with Intents

Example

Creating a LUIS Model requires little to no programming experience. However, you need to be familiar with 2 important terms that will be used extensively.

1. Intents - These are how you identify functions that need to be executed when the user types in something. Eg - An intent named Hi will identify a function that needs to be executed whenever the user sends "Hi". Intents are uniquely named in your program/model.
2. Entities - These identify the nouns in a statement. Eg - "Set an alarm for 1:00 pm", here 1:00 pm is an entity that needs to be recognized by the chat-bot to set an alarm.

Note: Images of the website are not provided as the front-end my change, but the core concept remains the same.

To create a new model, go to LUIS.ai and sign-in with your Microsoft Account to be taken to the app creation page. Where a blank project can be created.

Defining Intents:

Intents can be defined on the Intents tab. They identify what function you need to perform when the user enters anything.

All applications have a default None intent, which is activated whenever the user input matches no other intent.

To define an intent,

1. Give it a unique name relevant to the function you want to perform.
2. Once the naming is complete, you should add utterances to the intent. Utterances are what you want the user to send in order to activate the intent that you are defining. Try feeding as many different utterances as possible in order for LUIS to associate intents and utterances properly.
3. Train your LUIS Model, by clicking the Train button on Train and Test Tab. After training the app can be tested in the panel below.
4. Finally publish your app in the Publish App Tab. You should now get an endpoint URL that should be put in while defining LUISRecognizer in your bot code.