Train, test, publish, and review a conversational language understanding model

Completed

Creating a model is an iterative process with the following activities:

Diagram that shows the train, test, publish, review cycle.

  1. Train a model to learn intents and entities from sample utterances.
  2. Test the model interactively or using a testing dataset with known labels
  3. Deploy a trained model to a public endpoint so client apps can use it
  4. Review predictions and iterate on utterances to train your model

By following this iterative approach, you can improve the language model over time based on user input, helping you develop solutions that reflect the way users indicate their intents using natural language.