Microsoft Azure AI Fundamentals (AI-900) 5 — Questions and Answers
Question 1: Which Azure Machine Learning feature automatically tries multiple algorithms and hyperparameter combinations to find the best model for a dataset?
- Azure ML Pipelines
- Automated ML (AutoML) (Correct answer)
- Azure ML Endpoints
- Data labeling
Correct answer: Automated ML (AutoML)
Automated ML iterates through algorithm and feature preprocessing combinations, evaluating each using cross-validation to surface the best-performing model.
Question 2: What is 'responsible AI' most accurately described as?
- A set of Azure pricing tiers for AI services
- A framework of principles guiding the ethical and trustworthy development and use of AI (Correct answer)
- A certification required before deploying Azure AI models in production
- A monitoring tool that detects AI model failures
Correct answer: A framework of principles guiding the ethical and trustworthy development and use of AI
Responsible AI is a set of guiding principles—including fairness, reliability, privacy, inclusiveness, transparency, and accountability—for building trustworthy AI systems.
Question 3: In natural language processing, what does 'tokenization' refer to?
- Assigning sentiment scores to words
- Splitting text into individual units such as words or subwords for processing (Correct answer)
- Converting text to numerical vector embeddings
- Identifying the language in which a document is written
Correct answer: Splitting text into individual units such as words or subwords for processing
Tokenization breaks raw text into tokens (words, subwords, or characters) that serve as the basic units for downstream NLP processing.
Question 4: Which Azure service would you use to translate text between more than 100 languages in real time?
- Azure Language Understanding (LUIS)
- Azure Text Analytics
- Azure Translator (Correct answer)
- Azure Speech Service
Correct answer: Azure Translator
Azure Translator is a cloud-based neural machine translation service that supports text translation, transliteration, and language detection across 100+ languages.
Question 5: What is a 'confusion matrix' used for in machine learning evaluation?
- Visualizing the distribution of training data features
- Showing the counts of correct and incorrect predictions broken down by class (Correct answer)
- Displaying the correlation between input features
- Tracking model training loss over epochs
Correct answer: Showing the counts of correct and incorrect predictions broken down by class
A confusion matrix tabulates true positives, true negatives, false positives, and false negatives, giving a detailed view of classification errors.
Question 6: Which component of an Azure ML workspace stores the versioned datasets, models, and experiment run history?
- Compute instance
- Datastore and registry artifacts (Correct answer)
- Inference cluster
- Environment definition
Correct answer: Datastore and registry artifacts
Datastores link to Azure storage accounts to hold datasets, while the workspace registry tracks versioned models and experiment artifacts for reproducibility.
Question 7: A hospital wants to predict which patients are at high risk for readmission within 30 days. What type of machine learning problem is this?
- Clustering
- Regression
- Binary classification (Correct answer)
- Anomaly detection
Correct answer: Binary classification
Predicting whether a patient will be readmitted within 30 days is a binary classification task with two outcomes: readmitted or not readmitted.
Microsoft Azure AI Fundamentals (AI-900) 5
Which Azure Machine Learning feature automatically tries multiple algorithms and hyperparameter combinations to find the best model for a dataset?