Microsoft Certified: Azure AI Engineer AssociateImplement natural language processing solutionsHard
A developer is building a system that needs to transcribe short voice commands (e.g., 'open app', 'play music') from users in a noisy environment. The system requires high accuracy for these specific commands, even with background noise. Which Azure AI Speech feature allows for improving the accuracy of transcribing specific phrases or commands?
- ASpeaker Diarization
- BText-to-Speech (TTS)
- CSpeech Translation
- DCustom Speech (Phrase List or Language Model)
Show answer & explanationAnswer & explanation
Correct answer: D. Custom Speech (Phrase List or Language Model)
Custom Speech allows for tailoring the speech-to-text models to specific scenarios. By providing a Phrase List of expected commands or training a Custom Language Model with domain-specific vocabulary, the system can significantly improve the accuracy of transcribing those particular commands, even in challenging environments.
Why the other options are wrong
- A. Speaker Diarization separates speakers, not improves recognition of specific phrases.
- B. Text-to-Speech converts text to speech, not speech to text.
- C. Speech Translation converts spoken language between different languages, not improves transcription accuracy for specific phrases within a single language.
Azure Custom Speech
A set of features within Azure AI Speech that allows users to customize speech-to-text models by providing domain-specific data (e.g., audio, text, phrase lists) to improve transcription accuracy for specific scenarios or vocabulary.
- Improves speech-to-text accuracy.
- Uses phrase lists or custom language models.
- Beneficial for specialized vocabulary or noisy environments.
Memory trick: Custom Speech tunes the ear for tricky words.