Phase 4: Natural Language Processing (Questions 29-35)

Question 29: In speech synthesis, which process breaks text into individual words for phonetic assignment?

Rationale: Tokenization = breaking text into words. Transcription is speech-to-text. Lemmatization is word normalization.

Question 30: What confidence score does Azure Language detection return for an unknown language?

Rationale: Unknown language returns NaN (Not a Number) with language name "Unknown".

Question 31: Which TWO Azure AI Services enable text-to-text AND speech-to-text translation?

Rationale: Translator does text-to-text. Speech service does speech-to-text. Combined they cover both.

Question 32: Which Azure AI Service for Language feature provides online links for extracted entities?

Rationale: Entity linking connects entities to Wikipedia/knowledge base articles.

Question 33: Which Azure Speech feature identifies distinct user voices?

Rationale: Speaker recognition (part of speech recognition capabilities) identifies individual voices.

Question 34: Which type of translation does Azure Translator in Foundry Tools support?

Rationale: Azure Translator is text-to-text ONLY. Speech translation uses Azure Speech.

Question 35: Which THREE sources can generate questions and answers for a knowledge base?

Rationale: Valid sources: web pages, documents, manual entry. Audio and images cannot be directly imported.