5.3. Reflection Checkpoint: NLP Mastery

What breaks without mastering NLP concepts? The exam frequently tests service boundaries—which service handles text versus speech versus translation. Imagine seeing "real-time translation of spoken language" and choosing Azure Translator because it handles translation—wrong, Translator is text-only; speech translation requires Azure AI Speech.

Consider these questions like a service boundary test. For instance, the language detection question seems simple but tests whether you know the exact output format. If you hesitate, review Section 5.2. The Speech vs. Language distinction trips up many candidates.

1. Which NLP feature provides links to Wikipedia for further information about extracted entities?

   • Entity Linking. It disambiguates entities and connects them to knowledge bases.

2. Which Azure service supports text-to-text translation?

   • Azure Translator in Foundry Tools. It supports 60+ languages for text translation.

3. What three values does language detection return?

   • Language Name, ISO 639-1 Code, and Confidence Score.

4. Which feature of Azure AI Speech can identify distinct user voices?

   • Speech Recognition (specifically speaker recognition). It maps audio patterns to distinct voices.

5. A company needs to generate closed captions for live presentations. Which service should they use?

   • Azure AI Speech (Speech-to-Text). This converts spoken audio to text for captioning.