Tag: AI-900: Microsoft Azure AI Fundamentals

AI, AI-900, Artificial Intelligence (AI), Deep Learning, Microsoft Certification

Practice Exam Questions: Identify Features of the Transformer Architecture (AI-900 Exam Prep)

Practice Exam Questions

Question 1

What is the primary purpose of the self-attention mechanism in a Transformer model?

A. To reduce the size of the training dataset
B. To allow the model to focus on relevant parts of the input sequence
C. To replace the need for training data
D. To process words strictly in order

Correct Answer: B

Explanation:
Self-attention enables a Transformer to determine which words in a sentence are most relevant to one another, improving context understanding. It does not enforce strict order or reduce dataset size.

Question 2

Which feature allows Transformers to be trained more efficiently than recurrent neural networks (RNNs)?

A. Sequential word processing
B. Parallel processing of input data
C. Manual feature engineering
D. Rule-based language models

Correct Answer: B

Explanation:
Transformers process entire sequences in parallel, unlike RNNs that process tokens sequentially. This makes Transformers faster and more scalable.

Question 3

A key reason Transformers require positional encoding is because they:

A. Use convolutional layers
B. Process all input tokens at the same time
C. Rely on labeled data only
D. Perform unsupervised learning

Correct Answer: B

Explanation:
Because Transformers process words in parallel, positional encoding is needed to preserve information about word order in a sentence.

Question 4

Which type of AI workload most commonly uses Transformer-based models?

A. Time-series forecasting
B. Natural language processing
C. Image compression
D. Robotics control systems

Correct Answer: B

Explanation:
Transformers are primarily used for NLP tasks such as translation, summarization, and conversational AI.

Question 5

Which statement best describes the encoder–decoder architecture used in many Transformer models?

A. Both components generate output text
B. The encoder understands input, and the decoder generates output
C. The decoder trains the encoder
D. Both components store training data

Correct Answer: B

Explanation:
The encoder processes and understands the input sequence, while the decoder generates the output sequence based on that understanding.

Question 6

Why are Transformers better at handling long-range dependencies in text compared to earlier models?

A. They use fewer parameters
B. They rely on handcrafted grammar rules
C. They use attention to relate all words in a sequence
D. They process words one at a time

Correct Answer: C

Explanation:
Self-attention allows Transformers to evaluate relationships between all words in a sentence, regardless of distance.

Question 7

Which Azure scenario is most likely to involve a Transformer-based model?

A. Predicting tomorrow’s stock price
B. Detecting network hardware failures
C. Translating text between languages
D. Calculating average sales per region

Correct Answer: C

Explanation:
Language translation is a classic NLP task that relies heavily on Transformer architectures.

Question 8

What is a major advantage of Transformers over traditional sequence models?

A. They require no training data
B. They eliminate bias automatically
C. They improve scalability and performance
D. They work only with structured data

Correct Answer: C

Explanation:
Transformers scale efficiently due to parallel processing and attention mechanisms, improving performance on large datasets.

Question 9

Which statement about Transformers is TRUE?

A. They are rule-based AI systems
B. They process data strictly sequentially
C. They are a type of deep learning model
D. They are limited to image recognition

Correct Answer: C

Explanation:
Transformers are deep learning architectures commonly used for NLP tasks.

Question 10

Which feature enables a Transformer model to understand the context of a word based on surrounding words?

A. Positional encoding
B. Tokenization
C. Self-attention
D. Data labeling

Correct Answer: C

Explanation:
Self-attention allows the model to weigh the importance of surrounding words when interpreting meaning and context.

Quick Exam Tip

If you see keywords like:

  • attention
  • context
  • parallel processing
  • language understanding
  • Azure OpenAI

You’re almost certainly dealing with a Transformer-based model.

Go to the AI-900 Exam Prep Hub main page.

AI, AI-900, Artificial Intelligence (AI), Deep Learning, Microsoft Certification, Natural Language Processing (NLP)

Identify Features of the Transformer Architecture (AI-900 Exam Prep)

Where This Topic Fits in the Exam

  • Exam domain: Describe fundamental principles of machine learning on Azure (15–20%)
  • Sub-area: Identify common machine learning techniques
  • Focus: Understanding what Transformers are, why they matter, and what problems they solve — not how to code them

The AI-900 exam tests conceptual understanding, so you should recognize key features, benefits, and common use cases of the Transformer architecture.

What Is the Transformer Architecture?

The Transformer architecture is a type of deep learning model designed primarily for natural language processing (NLP) tasks.
It was introduced in the paper “Attention Is All You Need” and has since become the foundation for modern AI models such as:

  • Large Language Models (LLMs)
  • Chatbots
  • Translation systems
  • Text summarization tools

Unlike earlier sequence models, Transformers do not process data sequentially. Instead, they analyze entire sequences at once, which makes them faster and more scalable.

Key Features of the Transformer Architecture

1. Attention Mechanism (Self-Attention)

The core feature of a Transformer is self-attention.

Self-attention allows the model to:

  • Evaluate the importance of each word relative to every other word in a sentence
  • Understand context and relationships, even when words are far apart

Example:
In the sentence “The animal didn’t cross the road because it was tired”, self-attention helps the model understand what “it” refers to.

📌 Exam takeaway: Transformers use attention to understand context more effectively than older models.

2. Parallel Processing

Traditional models like RNNs process text one word at a time.
Transformers process all words in parallel.

Benefits:

  • Faster training
  • Better performance on large datasets
  • Improved scalability in cloud environments (like Azure)

📌 Exam takeaway: Transformers are efficient and scalable because they don’t rely on sequential processing.

3. Encoder–Decoder Structure

Many Transformer-based models use an encoder–decoder architecture:

  • Encoder:
    • Reads and understands the input (e.g., a sentence in English)
  • Decoder:
    • Generates the output (e.g., the translated sentence in Spanish)

📌 Exam takeaway: Transformers often use encoders to understand input and decoders to generate output.

4. Positional Encoding

Because Transformers process words in parallel, they need a way to understand word order.

Positional encoding:

  • Adds information about the position of each word
  • Allows the model to understand sentence structure and sequence

📌 Exam takeaway: Transformers use positional encoding to retain word order information.

5. Strong Performance on Natural Language Tasks

Transformers are especially effective for:

  • Text translation
  • Text summarization
  • Question answering
  • Chatbots and conversational AI
  • Sentiment analysis

📌 Exam takeaway: Transformers are closely associated with natural language processing workloads.

Why Transformers Are Important in Azure AI

Microsoft Azure AI services rely heavily on Transformer-based models, especially in:

  • Azure OpenAI Service
  • Azure AI Language
  • Conversational AI and copilots
  • Search and knowledge mining

Understanding Transformers helps explain why modern AI solutions are more accurate, context-aware, and scalable.

Transformers vs Earlier Models (High-Level)

FeatureEarlier Models (RNNs/CNNs)Transformers
Sequence processingSequentialParallel
Context handlingLimitedStrong
Long-range dependenciesDifficultEffective
Training speedSlowerFaster
NLP performanceModerateState-of-the-art

📌 Exam focus: You don’t need technical depth — just understand why Transformers are better for language tasks.

Common Exam Pitfalls to Avoid

  • ❌ Thinking Transformers replace all ML models
  • ❌ Assuming Transformers are only for images
  • ❌ Confusing Transformers with traditional rule-based NLP

✅ Remember: Transformers are deep learning models optimized for language and sequence understanding.

Key Exam Summary (Must-Know Points)

If you remember nothing else, remember this:

  • Transformers are deep learning models
  • They rely on self-attention
  • They process data in parallel
  • They are especially effective for natural language processing
  • They power modern AI services in Azure

Go to the Practice Exam Questions for this topic.

Go to the AI-900 Exam Prep Hub main page.

AI, AI-900, Artificial Intelligence (AI), Machine Learning (ML), Microsoft Certification

Identify Features and Labels in a Dataset for Machine Learning (AI-900 Exam Prep)

This section of the AI-900: Microsoft Azure AI Fundamentals exam focuses on understanding one of the most important foundational concepts in machine learning: features and labels. You are not expected to build models or write code, but you must be able to recognize features and labels in a dataset and understand their role in different machine learning scenarios.

This topic appears under: Describe Artificial Intelligence workloads and considerations (15–20%) → Describe core machine learning concepts

What Is a Dataset in Machine Learning?

A dataset is a collection of data used to train, validate, and test machine learning models. In supervised learning scenarios (which are emphasized in AI-900), a dataset typically contains:

  • Features: The input values used to make predictions
  • Labels: The output or target values the model learns to predict

Each row in a dataset usually represents a single observation or record, and each column represents either a feature or a label.

What Are Features?

Features are the individual measurable properties or characteristics of the data that are used as inputs to a machine learning model.

Key Characteristics of Features

  • Features describe what you know about each data point
  • They are used by the model to identify patterns
  • Features can be numerical, categorical, or derived

Examples of Features

ScenarioExample Features
House price predictionNumber of bedrooms, square footage, location
Customer churnAccount age, number of support tickets, monthly spend
Email classificationWord frequency, sender domain, message length

In Azure Machine Learning, features are often referred to as input variables.

What Are Labels?

A label is the value that a machine learning model is trained to predict. Labels are only present in supervised learning datasets.

Key Characteristics of Labels

  • Labels represent the outcome or answer
  • A dataset usually has one label column
  • Labels are known during training but unknown during prediction

Examples of Labels

ScenarioLabel
House price predictionSale price
Customer churnChurned (Yes/No)
Image classificationObject category

In Azure Machine Learning, labels are often called target variables.

Features vs Labels: Key Differences

AspectFeaturesLabels
PurposeInput to the modelOutput to predict
QuantityUsually manyTypically one
Known during trainingYesYes
Known during predictionYesNo

Understanding this distinction is critical for AI-900 exam questions.

Features and Labels in Supervised Learning

Supervised learning relies on labeled datasets. The model learns by comparing its predictions to the known labels and adjusting accordingly.

Common Supervised Learning Types

  • Regression
    • Features: numeric or categorical inputs
    • Label: numeric value (e.g., price, temperature)
  • Classification
    • Features: descriptive inputs
    • Label: category or class (e.g., spam/not spam)

Features and Labels in Unsupervised Learning

Unsupervised learning datasets do not contain labels.

  • The model identifies patterns or groupings on its own
  • Common example: clustering

In AI-900, this distinction is important:

If a dataset has no labels, it is not supervised learning.

Real-World Azure Example

Consider a dataset used in Azure Machine Learning to predict whether a customer will cancel a subscription.

  • Features:
    • Number of logins per month
    • Subscription length
    • Customer support interactions
  • Label:
    • Subscription canceled (Yes or No)

The model learns the relationship between the features and the label to make future predictions.

Exam Tips for AI-900

  • If the question asks “what the model uses to make predictions”, look for features
  • If the question asks “what the model predicts”, look for labels
  • If labels are present, it is supervised learning
  • AI-900 focuses on conceptual understanding, not data science implementation

Key Takeaways

  • Features are input variables used to make predictions
  • Labels are the known outcomes the model learns to predict
  • Supervised learning requires labeled data
  • Being able to identify features and labels in a scenario is essential for AI-900

This knowledge forms the foundation for understanding regression, classification, and many Azure AI workloads covered later in the exam.

Go to the Practice Exam Questions for this topic.

Go to the AI-900 Exam Prep Hub main page.

AI, AI-900, Artificial Intelligence (AI), Machine Learning (ML), Microsoft Certification

Practice Questions: Describe How Training and Validation Datasets Are Used in Machine Learning (AI-900 Exam Prep)

Practice Exam Questions

Question 1

What is the primary purpose of a training dataset in machine learning?

A. To evaluate the model’s accuracy on new data
B. To teach the model patterns using known outcomes
C. To store prediction results
D. To deploy the model to production

Correct Answer: B

Explanation:
The training dataset is used to teach the model by learning relationships between features and labels.

Question 2

Which dataset is used to assess how well a machine learning model performs on unseen data?

A. Training dataset
B. Feature dataset
C. Validation dataset
D. Prediction dataset

Correct Answer: C

Explanation:
The validation dataset is separate from training data and is used to evaluate the model’s ability to generalize.

Question 3

Why should the same dataset not be used for both training and validation?

A. It increases storage costs
B. It slows down training
C. It can lead to misleading performance results
D. It prevents model deployment

Correct Answer: C

Explanation:
Using the same data for training and validation can hide overfitting and give an inaccurate measure of model performance.

Question 4

A model performs very well on training data but poorly on validation data. What is this most likely an example of?

A. Underfitting
B. Overfitting
C. Data labeling
D. Feature engineering

Correct Answer: B

Explanation:
Overfitting occurs when a model memorizes training data but fails to generalize to new, unseen data.

Question 5

Which statement about a validation dataset is TRUE?

A. It is used to adjust model parameters
B. It replaces the need for training data
C. It helps evaluate model performance
D. It contains only unlabeled data

Correct Answer: C

Explanation:
Validation data is used to assess how well the model performs but is not used to train or adjust it.

Question 6

In supervised learning, which datasets typically contain both features and labels?

A. Validation only
B. Training only
C. Both training and validation
D. Neither training nor validation

Correct Answer: C

Explanation:
Both datasets contain features and labels, but they are used for different purposes.

Question 7

What is a key benefit of using a validation dataset during model development?

A. Faster training times
B. Automatic feature creation
C. Detection of overfitting
D. Reduced data storage

Correct Answer: C

Explanation:
Validation data helps identify whether the model is overfitting the training data.

Question 8

A dataset is split into 80% training data and 20% validation data.
What is the purpose of the 20% portion?

A. To retrain the model after deployment
B. To evaluate the model’s predictions
C. To generate new features
D. To label the data

Correct Answer: B

Explanation:
The validation portion is used to evaluate how well the model performs on unseen data.

Question 9

Which phrase best describes how a validation dataset is used?

A. Teaching the model
B. Fine-tuning the labels
C. Testing model generalization
D. Storing predictions

Correct Answer: C

Explanation:
Validation data is used to test how well the model generalizes beyond its training data.

Question 10

Which scenario correctly describes the use of training and validation datasets?

A. Training data is used only after deployment
B. Validation data is used to adjust model weights
C. Training data teaches the model; validation data evaluates it
D. Both datasets are identical

Correct Answer: C

Explanation:
Training data is used for learning, while validation data is used for evaluation.

Exam Strategy Tip

On AI-900:

  • Training dataset → learning and pattern recognition
  • Validation dataset → evaluation and generalization
  • Watch for keywords like overfitting, unseen data, and model performance

If you can map those keywords quickly, these questions become easy points.

Go to the AI-900 Exam Prep Hub main page.

AI, AI-900, Artificial Intelligence (AI), Machine Learning (ML), Microsoft Certification

Describe How Training and Validation Datasets Are Used in Machine Learning (AI-900 Exam Prep)

This section of the AI-900: Microsoft Azure AI Fundamentals exam focuses on understanding how machine learning models learn from data and how their performance is evaluated. Specifically, it covers the role of training datasets and validation datasets, which are core concepts in supervised machine learning.

This topic appears under: Describe fundamental principles of machine learning on Azure (15–20%) → Describe core machine learning concepts

You are not expected to build or tune models for AI-900, but you must be able to describe the purpose of training and validation datasets and how they differ.

Why Datasets Are Split in Machine Learning

In machine learning, using the same data to both train and evaluate a model can lead to misleading results. To avoid this, datasets are commonly split into separate subsets, each with a distinct purpose.

At a minimum, most machine learning workflows use:

  • A training dataset
  • A validation dataset

These datasets help ensure that a model can generalize to new, unseen data.

Training Dataset

A training dataset is the portion of data used to teach the machine learning model how to make predictions.

Key Characteristics of Training Data

  • Contains both features and labels (in supervised learning)
  • Used to identify patterns and relationships in the data
  • Typically makes up the largest portion of the dataset

What Happens During Training

  • The model makes predictions using the features
  • Predictions are compared to the known labels
  • The model adjusts its internal parameters to reduce errors

In Azure Machine Learning, this is the phase where the model “learns” from historical data.

Validation Dataset

A validation dataset is used to evaluate how well the model performs on unseen data during the training process.

Key Characteristics of Validation Data

  • Separate from the training dataset
  • Contains features and labels
  • Used to assess model accuracy and generalization

Why Validation Data Is Important

  • Helps detect overfitting (when a model memorizes training data)
  • Provides an unbiased evaluation of model performance
  • Supports decisions about model selection or improvement

For AI-900, the key idea is that validation data is not used to train the model, only to evaluate it.

Training vs Validation: Key Differences

AspectTraining DatasetValidation Dataset
Primary purposeTeach the modelEvaluate the model
Used to adjust model parametersYesNo
Seen by the model during learningYesNo
Helps detect overfittingIndirectlyYes

Understanding this distinction is essential for AI-900 exam questions.

Common Data Split Ratios

While AI-900 does not test exact percentages, common industry practices include:

  • 70% training / 30% validation
  • 80% training / 20% validation

The exact split depends on dataset size and use case, but the concept is what matters for the exam.

Example Scenario

A company is building a model to predict whether customers will cancel a subscription.

  • Training dataset:
    • Used to teach the model using historical customer behavior and known outcomes
  • Validation dataset:
    • Used to test how accurately the model predicts cancellations for customers it has not seen before

This approach helps ensure the model performs well in real-world scenarios.

Overfitting and Generalization

One of the main reasons for using a validation dataset is to avoid overfitting.

  • Overfitting occurs when a model performs well on training data but poorly on new data
  • Validation data helps confirm that the model can generalize beyond the training set

For AI-900, you only need to recognize this relationship, not the mathematical details.

Azure Context for AI-900

In Azure Machine Learning:

  • Training data is used to train machine learning models
  • Validation data is used to evaluate model performance during development
  • This separation supports reliable and responsible AI solutions

Exam Tips for AI-900

  • If the question mentions learning or adjusting the model, think training dataset
  • If the question mentions evaluation or performance on unseen data, think validation dataset
  • Validation data is not used to teach the model
  • AI-900 focuses on understanding why datasets are separated

Key Takeaways

  • Training datasets are used to teach machine learning models
  • Validation datasets are used to evaluate model performance
  • Separating datasets helps prevent overfitting
  • Understanding these roles is a core AI-900 exam skill

Go to the Practice Exam Questions for this topic.

Go to the AI-900 Exam Prep Hub main page.

AI, AI-900, Artificial Intelligence (AI), Machine Learning (ML), Microsoft Certification

Practice Questions: Describe Capabilities of Automated Machine Learning (AI-900 Exam Prep)

Practice Exam Questions

Question 1

What is the primary purpose of Automated Machine Learning (AutoML) in Azure?

A. To replace data scientists
B. To automatically label data
C. To select and optimize machine learning models
D. To deploy models without evaluation

Correct Answer: C

Explanation:
AutoML automatically selects algorithms and tunes parameters to identify the best-performing model for a given dataset.

Question 2

Which machine learning scenarios are supported by Azure Automated Machine Learning?

A. Clustering only
B. Regression and classification
C. Reinforcement learning
D. Rule-based automation

Correct Answer: B

Explanation:
AutoML supports supervised learning scenarios such as regression and classification, which are core to AI-900.

Question 3

How does AutoML reduce the need for deep machine learning expertise?

A. By eliminating the need for training data
B. By automatically selecting models and hyperparameters
C. By generating business requirements
D. By replacing human oversight

Correct Answer: B

Explanation:
AutoML handles model selection and hyperparameter tuning automatically, reducing manual effort and expertise requirements.

Question 4

Which task is handled automatically by Azure AutoML?

A. Defining business objectives
B. Cleaning poor-quality data
C. Hyperparameter tuning
D. Approving model deployment

Correct Answer: C

Explanation:
AutoML automatically adjusts hyperparameters to improve model performance.

Question 5

A team wants to quickly build a sales forecasting model with minimal manual configuration.
Which Azure capability should they use?

A. Azure Cognitive Services
B. Azure Bot Service
C. Automated Machine Learning
D. Azure Logic Apps

Correct Answer: C

Explanation:
AutoML is designed to quickly build supervised ML models, including time-series forecasting.

Question 6

Which statement about Automated Machine Learning is TRUE?

A. AutoML guarantees perfect model accuracy
B. AutoML removes the need for human review
C. AutoML compares multiple models automatically
D. AutoML works only with unlabeled data

Correct Answer: C

Explanation:
AutoML evaluates and compares multiple models to identify the best-performing option.

Question 7

Which Azure service provides Automated Machine Learning capabilities?

A. Azure Functions
B. Azure Machine Learning
C. Azure App Service
D. Azure Synapse Analytics

Correct Answer: B

Explanation:
Automated Machine Learning is a feature within Azure Machine Learning.

Question 8

What is a key benefit of using AutoML?

A. Manual feature engineering
B. Faster model development
C. Elimination of data preparation
D. Guaranteed regulatory compliance

Correct Answer: B

Explanation:
AutoML speeds up model development by automating model selection, tuning, and evaluation.

Question 9

Which of the following is NOT a capability of Automated Machine Learning?

A. Automatic model evaluation
B. Automatic algorithm selection
C. Automatic business decision-making
D. Hyperparameter tuning

Correct Answer: C

Explanation:
AutoML supports model creation and evaluation but does not make business decisions.

Question 10

Why is Automated Machine Learning especially useful for beginners?

A. It removes the need for labeled data
B. It eliminates model deployment steps
C. It simplifies model creation and experimentation
D. It replaces Azure Machine Learning

Correct Answer: C

Explanation:
AutoML simplifies experimentation by automating many steps involved in building machine learning models.

Exam Strategy Tip

On AI-900, think of AutoML as a productivity accelerator:

  • You provide the data and goal
  • AutoML handles model selection, tuning, and evaluation
  • Humans still review and deploy the model

If a question mentions automatic selection, minimal configuration, or quick model building, the answer is might be related to Automated Machine Learning.

Go to the AI-900 Exam Prep Hub main page.

AI, AI-900, Artificial Intelligence (AI), Machine Learning (ML), Microsoft Certification

Describe Capabilities of Automated Machine Learning (AI-900 Exam Prep)

This section of the AI-900: Microsoft Azure AI Fundamentals exam focuses on understanding what Automated Machine Learning (AutoML) is and what it can do within Azure Machine Learning. The emphasis is on recognizing capabilities and use cases, not on configuring pipelines or writing code.

This topic appears under: Describe fundamental principles of machine learning on Azure (15–20%) → Describe Azure Machine Learning capabilities

What Is Automated Machine Learning?

Automated Machine Learning (AutoML) is a capability in Azure Machine Learning that automatically selects the best machine learning model and tuning settings for a given dataset and problem.

AutoML helps users:

  • Build machine learning models faster
  • Reduce the need for deep data science expertise
  • Focus on business problems rather than algorithms

For AI-900, you only need to understand what AutoML does, not how to implement it.

Problems AutoML Can Solve

Automated Machine Learning in Azure supports common supervised learning scenarios:

  • Regression – Predicting numeric values (for example, sales forecasts)
  • Classification – Predicting categories or classes (for example, fraud detection)
  • Time-series forecasting – Predicting values over time (for example, demand prediction)

AutoML does not focus on unsupervised learning scenarios such as clustering for AI-900.

Key Capabilities of Automated Machine Learning

Automatic Model Selection

AutoML automatically:

  • Tries multiple machine learning algorithms
  • Compares model performance
  • Selects the best-performing model based on evaluation metrics

Users do not need to manually choose algorithms.

Automated Hyperparameter Tuning

AutoML automatically adjusts hyperparameters to improve model performance, such as:

  • Learning rate
  • Number of trees
  • Regularization settings

This removes the need for manual trial-and-error tuning.

Built-in Feature Engineering

AutoML can automatically create and transform features, including:

  • Normalizing numeric data
  • Encoding categorical values
  • Handling missing values

This simplifies data preparation for machine learning.

Model Evaluation and Comparison

AutoML evaluates models using validation data and metrics such as:

  • Accuracy
  • Precision and recall
  • Mean absolute error

It then ranks models so users can easily compare results.

Integration with Azure Machine Learning

AutoML is fully integrated into Azure Machine Learning, allowing users to:

  • Track experiments
  • View model performance
  • Deploy selected models

This integration supports repeatable and responsible ML workflows.

Example Scenario

A retail company wants to predict monthly product sales but does not have a data science team.

Using Automated Machine Learning:

  • The company provides historical sales data
  • AutoML tests multiple regression models
  • The best-performing model is automatically selected

This allows faster model creation with minimal manual effort.

What AutoML Does NOT Do (Exam-Relevant)

It is important to recognize AutoML limitations for AI-900:

  • It does not eliminate the need for quality data
  • It does not automatically define business goals
  • It does not replace human oversight

AutoML assists model creation but does not remove responsibility from users.

Azure Context for AI-900

In Azure Machine Learning, AutoML:

  • Simplifies model creation
  • Supports beginners and non-experts
  • Accelerates experimentation and deployment

AI-900 questions often focus on why AutoML is useful rather than how it works internally.

Exam Tips for AI-900

  • If the question mentions automatic model selection or tuning, think AutoML
  • AutoML is best for quickly building supervised ML models
  • Remember: AutoML helps choose models, but humans still provide data and goals

Key Takeaways

  • Automated Machine Learning automates model selection, tuning, and evaluation
  • It supports regression, classification, and forecasting scenarios
  • AutoML reduces the need for deep ML expertise
  • Understanding its capabilities is essential for AI-900

This topic connects directly to Azure Machine Learning services and helps bridge core ML concepts with real-world Azure AI capabilities.

Go to the Practice Exam Questions for this topic.

Go to the AI-900 Exam Prep Hub main page.

AI, AI-900, Artificial Intelligence (AI), Data Science, Machine Learning (ML), Microsoft Certification

Practice Questions: Describe data and compute services for data science and machine learning (AI-900 Exam Prep)

Practice Exam Questions

Question 1

Which Azure service is most commonly used to store large, unstructured datasets for machine learning training?

A. Azure SQL Database
B. Azure Blob Storage
C. Azure Cosmos DB
D. Azure Virtual Machines

✅ Correct Answer: B. Azure Blob Storage

Explanation:
Azure Blob Storage is designed to store large amounts of unstructured data such as files, images, and CSVs. It is the most common data storage service used in machine learning workflows.

Question 2

Which Azure service is specifically designed to train, manage, and deploy machine learning models?

A. Azure Kubernetes Service (AKS)
B. Azure Machine Learning
C. Azure Data Factory
D. Azure App Service

✅ Correct Answer: B. Azure Machine Learning

Explanation:
Azure Machine Learning provides managed tools and compute for training, evaluating, and deploying machine learning models. It is the core ML service in Azure.

Question 3

You need to store structured, relational data that will be used to train a machine learning model. Which Azure service is most appropriate?

A. Azure Blob Storage
B. Azure Data Lake Storage
C. Azure SQL Database
D. Azure File Storage

✅ Correct Answer: C. Azure SQL Database

Explanation:
Azure SQL Database is used for structured data stored in tables with defined schemas, making it suitable for relational datasets used in machine learning.

Question 4

Which Azure service is primarily used to deploy machine learning models for scalable, real-time predictions?

A. Azure Virtual Machines
B. Azure Machine Learning compute
C. Azure Kubernetes Service (AKS)
D. Azure Blob Storage

✅ Correct Answer: C. Azure Kubernetes Service (AKS)

Explanation:
AKS is commonly used to deploy machine learning models in production environments where scalability and high availability are required.

Question 5

What is the primary purpose of compute resources in machine learning?

A. To store training data
B. To visualize data
C. To train and run machine learning models
D. To manage user access

✅ Correct Answer: C. To train and run machine learning models

Explanation:
Compute resources provide the processing power required to train models and perform inference.

Question 6

Which Azure service provides customizable compute environments, including GPU-based machines, for machine learning workloads?

A. Azure Functions
B. Azure Virtual Machines
C. Azure Logic Apps
D. Azure SQL Database

✅ Correct Answer: B. Azure Virtual Machines

Explanation:
Azure Virtual Machines allow users to fully control the operating system, software, and hardware configuration, making them ideal for specialized ML workloads.

Question 7

Which data service is best suited for big data analytics and large-scale machine learning workloads?

A. Azure Blob Storage
B. Azure SQL Database
C. Azure Data Lake Storage Gen2
D. Azure Table Storage

✅ Correct Answer: C. Azure Data Lake Storage Gen2

Explanation:
Azure Data Lake Storage Gen2 is optimized for analytics and big data workloads, making it ideal for large-scale machine learning scenarios.

Question 8

In a typical Azure machine learning workflow, where are trained models and output artifacts often stored?

A. Azure Virtual Machines
B. Azure Blob Storage
C. Azure SQL Database
D. Azure Active Directory

✅ Correct Answer: B. Azure Blob Storage

Explanation:
Blob Storage is commonly used to store trained models, logs, and experiment outputs due to its scalability and cost efficiency.

Question 9

Which Azure service combines data storage and analytics capabilities for machine learning and data science?

A. Azure Data Lake Storage
B. Azure File Storage
C. Azure App Service
D. Azure Functions

✅ Correct Answer: A. Azure Data Lake Storage

Explanation:
Azure Data Lake Storage is built for analytics and integrates well with data science and machine learning workloads.

Question 10

Which statement best describes Azure Machine Learning compute?

A. It is used only for storing machine learning data
B. It provides managed compute resources for training and inference
C. It replaces Azure Virtual Machines
D. It is used only for model deployment

✅ Correct Answer: B. It provides managed compute resources for training and inference

Explanation:
Azure Machine Learning compute offers scalable, managed CPU and GPU resources specifically designed for training and running machine learning models.

Final Exam Tips 🔑

For AI-900, remember these high-yield associations:

  • Blob Storage → unstructured ML data
  • Data Lake Storage → big data & analytics
  • Azure SQL Database → structured data
  • Azure Machine Learning → training & managing models
  • Virtual Machines → custom ML environments
  • AKS → scalable deployment

Go to the AI-900 Exam Prep Hub main page.

AI, AI-900, Artificial Intelligence (AI), Data Science, Machine Learning (ML)

Describe Data and Compute Services for Data Science and Machine Learning (AI-900 Exam Prep)

This topic focuses on understanding which Azure services are used to store data and provide compute power for data science and machine learning workloads — not on how to configure them in depth. For the AI-900 exam, you should recognize what each service is used for and when you would choose one over another.

Why Data and Compute Matter in Machine Learning

Machine learning solutions require two essential components:

  • Data services → where training and inference data is stored and accessed
  • Compute services → where models are trained and executed

Azure provides scalable, cloud-based services for both, allowing organizations to build, train, and deploy machine learning solutions efficiently.

Data Services for Machine Learning on Azure

Azure offers several data storage services commonly used in machine learning scenarios.

Azure Blob Storage

Azure Blob Storage is the most common data store for machine learning.

Key characteristics:

  • Stores unstructured data (files, images, videos, CSVs)
  • Highly scalable and cost-effective
  • Frequently used as the data source for Azure Machine Learning experiments

Typical use cases:

  • Training datasets
  • Model artifacts
  • Logs and output files

👉 On AI-900: If the question mentions large datasets, files, or unstructured data, Blob Storage is usually the answer.

Azure Data Lake Storage Gen2

Azure Data Lake Storage is optimized for big data analytics and machine learning.

Key characteristics:

  • Built on Azure Blob Storage
  • Supports hierarchical namespaces
  • Designed for analytics workloads

Typical use cases:

  • Large-scale machine learning projects
  • Advanced analytics and data science pipelines

👉 On AI-900: Think of Data Lake Storage when big data and analytics are mentioned.

Azure SQL Database

Azure SQL Database stores structured, relational data.

Key characteristics:

  • Table-based storage
  • Uses SQL for querying
  • Suitable for well-defined schemas

Typical use cases:

  • Business and transactional data
  • Structured datasets used in ML training

👉 On AI-900: If the data is relational and structured, Azure SQL Database is a common choice.

Compute Services for Machine Learning on Azure

Compute services provide the processing power needed to train and run machine learning models.

Azure Machine Learning Compute

Azure Machine Learning provides managed compute resources specifically designed for ML workloads.

Key characteristics:

  • Scalable CPU and GPU compute
  • Used for training and inference
  • Managed through Azure Machine Learning workspace

Typical use cases:

  • Model training
  • Experimentation
  • Batch inference

👉 On AI-900: This is the primary compute service for machine learning.

Azure Virtual Machines

Azure Virtual Machines (VMs) offer full control over the compute environment.

Key characteristics:

  • Customizable CPU or GPU configurations
  • Supports specialized ML workloads
  • More management responsibility

Typical use cases:

  • Custom machine learning environments
  • Legacy or specialized ML tools

👉 On AI-900: VMs appear when flexibility or custom configuration is required.

Azure Kubernetes Service (AKS)

AKS is used primarily for deploying machine learning models at scale.

Key characteristics:

  • Container orchestration
  • High availability and scalability
  • Often used for real-time inference

Typical use cases:

  • Production ML model deployment
  • Scalable inference endpoints

👉 On AI-900: AKS is associated with deployment, not training.

How These Services Work Together

In a typical Azure machine learning workflow:

  1. Data is stored in Blob Storage, Data Lake, or SQL Database
  2. Models are trained using Azure Machine Learning compute or VMs
  3. Models are deployed using Azure Machine Learning or AKS
  4. Predictions are generated and consumed by applications

Azure handles scalability, security, and integration across these services.

Key Exam Takeaways

For AI-900, remember:

  • Blob Storage → unstructured ML data
  • Data Lake Storage → big data analytics
  • Azure SQL Database → structured data
  • Azure Machine Learning compute → training and experimentation
  • Virtual Machines → custom compute environments
  • AKS → scalable model deployment

You are not expected to configure these services — only recognize their purpose.

Exam Tip 💡

If a question asks:

  • “Where is ML data stored?” → Blob Storage or Data Lake
  • “Where is the model trained?” → Azure Machine Learning compute
  • “How is a model deployed at scale?” → AKS

Go to the Practice Exam Questions for this topic.

Go to the AI-900 Exam Prep Hub main page.

AI, AI-900, Artificial Intelligence (AI), Machine Learning (ML), Microsoft Certification

Practice Questions: Describe Model Management and Deployment Capabilities in Azure Machine Learning (AI-900 Exam Prep)

Practice Questions

Question 1

You train multiple machine learning models using different algorithms and want to store them in a central location with version tracking. Which Azure Machine Learning capability should you use?

A. Azure Kubernetes Service
B. Model registration
C. Batch endpoints
D. Automated machine learning

Correct Answer: B

Explanation:
Model registration stores trained models in Azure Machine Learning, enabling centralized management, versioning, and reuse.

Question 2

Why is model versioning important in Azure Machine Learning?

A. To reduce compute costs
B. To allow rollback to previous model versions
C. To encrypt model files
D. To improve model accuracy automatically

Correct Answer: B

Explanation:
Model versioning allows teams to track changes over time and revert to earlier versions if newer models perform poorly.

Question 3

Which deployment option should you use when predictions must be returned immediately to a web application?

A. Batch endpoint
B. Training pipeline
C. Real-time endpoint
D. Experiment run

Correct Answer: C

Explanation:
Real-time endpoints provide low-latency predictions through REST APIs, making them suitable for applications that need immediate responses.

Question 4

A data science team wants to score millions of records overnight without requiring instant responses. Which deployment approach is most appropriate?

A. Real-time endpoint
B. Batch endpoint
C. Azure Functions
D. Model registration

Correct Answer: B

Explanation:
Batch endpoints are designed for large-scale, offline predictions and do not require low-latency responses.

Question 5

Which Azure Machine Learning feature tracks metrics, parameters, and outputs from training runs?

A. Model deployment
B. Experiment tracking
C. Model endpoint
D. Azure Blob Storage

Correct Answer: B

Explanation:
Experiment tracking captures training details such as metrics and parameters, enabling comparison and reproducibility.

Question 6

After training a model, what is the primary purpose of registering it in Azure Machine Learning?

A. To retrain the model automatically
B. To expose the model as an API
C. To store and manage the model for future use
D. To encrypt the dataset

Correct Answer: C

Explanation:
Registering a model allows it to be stored, versioned, and managed, making it available for deployment and reuse.

Question 7

Which Azure Machine Learning capability simplifies scaling and infrastructure management when deploying models?

A. Model versioning
B. Containerized deployment
C. Experiment tracking
D. Data labeling

Correct Answer: B

Explanation:
Azure Machine Learning packages models into containers, simplifying deployment, scaling, and infrastructure management.

Question 8

What is a key difference between real-time endpoints and batch endpoints?

A. Real-time endpoints do not require models
B. Batch endpoints are used only for training
C. Real-time endpoints provide immediate predictions
D. Batch endpoints use more accurate models

Correct Answer: C

Explanation:
Real-time endpoints return predictions immediately, while batch endpoints process large datasets asynchronously.

Question 9

Which task is part of model management rather than model deployment?

A. Exposing a REST API
B. Scaling compute resources
C. Registering and versioning models
D. Handling prediction requests

Correct Answer: C

Explanation:
Registering and versioning models are model management tasks. Deployment focuses on making models available for predictions.

Question 10

Which statement best describes Azure Machine Learning’s role in model deployment?

A. It requires manual server configuration
B. It automates model training only
C. It simplifies deploying models to scalable endpoints
D. It replaces Azure Kubernetes Service

Correct Answer: C

Explanation:
Azure Machine Learning abstracts infrastructure complexity, making it easier to deploy models as scalable endpoints.

Final Exam Tips ✅

  • Model registration = storage + versioning
  • Real-time endpoint = immediate predictions
  • Batch endpoint = large-scale, offline predictions
  • AI-900 tests concepts, not implementation details

Go to the AI-900 Exam Prep Hub main page.