AI-102 Question 220
Single answerYou are developing a voice-controlled application for a customer service system using Azure AI. The application should recognize specific keywords like 'help', 'cancel', and 'support' to trigger specific workflows. You decide to implement keyword recognition using Azure Cognitive Services. Which approach should you choose to achieve this functionality?
- A
Use the Speech SDK to configure a KeywordRecognizer with a predefined keyword model
- B
Create a custom Speech-to-Text model using Azure Custom Speech and manually extract keywords from the transcribed text
- C
Deploy an Azure Form Recognizer model to extract keywords from the user's voice data
- D
Implement Azure Text Analytics with the Key Phrase Extraction feature to identify keywords in real-time
Show answer and explanation
Correct answer: A
Explanation
The KeywordRecognizer in the Speech SDK is specifically designed for scenarios where predefined keywords need to be detected in real-time from voice input. It provides an efficient and accurate solution compared to other approaches, which are either not designed for real-time voice data or require additional processing steps to achieve similar functionality.
- A. Correct.
This is the correct approach. The Speech SDK provides a KeywordRecognizer feature that is specifically designed for recognizing predefined keywords in real-time voice input. This is ideal for scenarios where specific trigger words need to be detected.
- B. Incorrect.
While a custom Speech-to-Text model can transcribe audio input into text, it does not provide a direct mechanism for real-time keyword recognition. Additional processing would be required, making this approach less efficient.
- C. Incorrect.
Azure Form Recognizer is designed for extracting information from documents and forms, and it is not suitable for processing voice data or recognizing keywords.
- D. Incorrect.
Azure Text Analytics' Key Phrase Extraction is designed for analyzing text data and extracting key phrases, but it does not work with real-time voice input or provide keyword recognition capabilities.