AI-900 Practice Question: Describe features of Natural Language Processing workloads on Azure

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Translating medical terminology between different languages for international patients

  Why it's wrong here

  Azure AI Translator is a separate service designed to convert text between languages; it has no capability to identify clinical entities like diagnoses or medications. Text Analytics for Health, by contrast, performs domain-specific entity extraction and normalization (e.g., to UMLS/ICD-10) from English clinical notes. Therefore, translation addresses language barriers, not the information-extraction task described in the question.

• ✓

  Extracting diagnoses, medications, symptoms, procedures, and lab results from clinical text

  Why this is correct

  Text Analytics for Health (TA4H) is purpose-built to extract medical entities—diagnoses, medications, symptoms, procedures, and lab results—from unstructured clinical text. It also identifies entity relations and attribute assertions (e.g., negation, temporality) and links entities to standard vocabularies like UMLS and ICD-10-CM. This allows clinical NLP applications to query and analyze notes without needing custom model training, exactly matching the capability described in the question.

• ✗

  Generating medical reports from structured patient data in an EHR system

  Why it's wrong here

  Generating medical reports from structured EHR data is natural language generation, not healthcare NLP extraction. Text Analytics for Health consumes unstructured text (e.g., physician notes, discharge summaries) and outputs structured data, so applying it to already-structured data would be conceptually backwards. The service does not synthesize or author prose; it parses free text to surface meaningful medical entities and relationships.

• ✗

  Diagnosing patient conditions from their described symptoms using AI

  Why it's wrong here

  Text Analytics for Health is not a diagnostic engine; it lacks clinical reasoning that weighs evidence, differentials, and patient history. While it can extract symptom and condition entities, it does not infer a diagnosis from those entities or provide treatment recommendations. Autonomous symptom-based diagnosis is a high-risk AI application outside the scope of this NLP feature, which only structures the information present in the text.

Go deeper

Related to this question