Demo4

M05 DEMO #4

• Speech Integration

This code is provided for demo purposes only for course AI-102.

Requirements

• Azure Subscription
• C# dotnet core
• VS code

Speech Integration

1. The code based on the hands-on example converted to Node.js and adopted to execute on RPi. The sample is working with a microphone and expects input from the microphone of any sentence and recognizes it as text and evaluates it with the LUIS model. Finally provide the requested response of the time or date depending on location.

2. Use the import option to restore the LUIS model from file clock.lu.

3. Settings file appsettings.json should be updated based on your values.

   {
   "LSPredictionEndpoint": "Your LUIS endpoint",
   "LSPredictionKey": "Your LUIS key"
   }

4. When the application is starting it is expecting input from the microphone in one of the following sentences.

   • What time is it?
   • What date is it now?
   • What time in London?

5. In the output you can observe JSON responses from LUIS. You can re-train the model to fix unrecognized intents or locations.

   

The sample includes Speech recognition and Speech synthesizing code examples.