AI-102 Question 276
Select 3You are building a chatbot using Azure Language Understanding (LUIS) for a customer support application. The initial model performs well for basic queries but struggles with more complex or ambiguous user inputs. Which steps should you take to optimize the language understanding model for better performance?
- A
Review and add representative utterances to improve intent recognition.
- B
Reduce the number of intents to a single intent for simplicity.
- C
Use Active Learning to identify and label low-confidence utterances.
- D
Enable and configure the Sentiment Analysis feature in the LUIS portal.
- E
Retrain the model after adding new examples to improve accuracy.
Show answer and explanation
Correct answers: A, C, E
Explanation
Optimizing a LUIS model involves iterative improvements through adding representative utterances, leveraging Active Learning to label low-confidence predictions, and retraining the model with updated data. These steps ensure the model adapts to diverse user inputs and improves its accuracy over time. Reducing intents or enabling unrelated features like Sentiment Analysis will not address the core issue of intent misclassification or ambiguity.
- A. Correct.
Adding representative utterances helps the model better understand diverse ways that users might phrase their queries, improving intent recognition.
- B. Incorrect.
Reducing the number of intents to a single intent oversimplifies the model and defeats the purpose of intent classification, reducing its usefulness.
- C. Correct.
Active Learning is a key strategy for identifying utterances where the model is uncertain, allowing you to label them and refine the model.
- D. Incorrect.
While Sentiment Analysis is a useful feature, it does not directly optimize the intent recognition capabilities of your LUIS model.
- E. Correct.
Retraining the model after adding new examples is essential for the model to learn from the updated data and improve its performance.