AI-102 Question 261
Select 3You are designing a conversational AI solution using Azure Language Understanding (LUIS). You need to create a new intent to handle customer queries about order statuses. Which actions should you take to ensure the intent can accurately distinguish between different user queries?
- A
Add multiple example utterances that represent different ways users might ask about order statuses.
- B
Use a prebuilt intent from LUIS for order status queries and avoid creating a custom intent.
- C
Ensure the utterances for the intent are diverse and cover a wide range of possible phrasings.
- D
Add example utterances that are highly similar to each other to improve intent recognition.
- E
Test the intent with sample queries after adding utterances to validate its accuracy.
Show answer and explanation
Correct answers: A, C, E
Explanation
To create an effective intent in LUIS, you need to provide diverse and representative utterances that cover a wide range of possible user queries. This improves the model's ability to generalize and distinguish between intents. Testing the intent with sample queries is also essential to validate the model's accuracy and identify areas for improvement. Using prebuilt intents might not be suitable for custom scenarios, and adding highly similar utterances does not enhance the model's performance.
- A. Correct.
Adding multiple example utterances helps LUIS understand the different ways users might phrase their queries, improving intent recognition.
- B. Incorrect.
While prebuilt intents in LUIS are useful, they may not always match your specific requirements. For a custom scenario like order status, creating a custom intent is generally more appropriate.
- C. Correct.
Ensuring diversity in the utterances helps LUIS generalize better to unseen queries and prevents overfitting to specific phrasings.
- D. Incorrect.
Adding highly similar utterances does not improve the model's ability to generalize. Instead, it may lead to redundancy and limited understanding of diverse phrasing.
- E. Correct.
Testing the intent with sample queries ensures that the model performs as expected and allows you to identify and address any shortcomings in the training data.