AI-900: Microsoft Azure AI Fundamentals

Sample Questions
Last updated: 1/19/2022
PLEASE COMPLETE THIS SURVEY (https://aka.ms/samplequestions)
Microsoft is exploring the possibility of providing sample questions as an exam preparation resource,
and we would like your feedback. While we prefer that you complete the survey after taking the exam,
you may complete it at any time. Thank You!

User Guide
These sample questions are intended to provide an overview of the style, wording, and difficulty of the
questions that you are likely to experience on this exam. These questions are not the same as what you
will see on the exam nor is this document illustrative of the length of the exam or its complexity (e.g.,
you may see additional question types, multiple case studies, and possibly labs). These questions are
examples only to provide insight into what to expect on the exam and help you determine if additional
preparation is required.
In the first section, you will find the questions without answers so that you can test your knowledge. In
the second section, the answer, a rationale, and a URL that will link you to additional information is
provided immediately below each question.

Contents
Questions-------------------------------------------------------------------------------------------------------------------------- 3
Question # 1 (Multiple Choice) -------------------------------------------------------------------------------------------- 3
Question # 2 (Matching)----------------------------------------------------------------------------------------------------- 3
Question # 3 (Sentence completion) ------------------------------------------------------------------------------------- 3
Question # 4 (Sentence completion) ------------------------------------------------------------------------------------- 3
Question # 5 (Multiple Choice) -------------------------------------------------------------------------------------------- 4
Question # 6 (Matching)----------------------------------------------------------------------------------------------------- 4
Question # 7 (Sentence completion) ------------------------------------------------------------------------------------- 4
Question # 8 (Multiple Choice) -------------------------------------------------------------------------------------------- 4
Question # 9 (Sentence completion) ------------------------------------------------------------------------------------- 5
Question # 10 (Multiple Choice)------------------------------------------------------------------------------------------- 5
Question # 11 (Multiple Choice)------------------------------------------------------------------------------------------- 5
Question # 12 (Sentence completion) ----------------------------------------------------------------------------------- 6
Question # 13 (Multiple Choice)------------------------------------------------------------------------------------------- 6
Question # 14 (Multiple Choice)------------------------------------------------------------------------------------------- 6
Question # 15 (Matching) --------------------------------------------------------------------------------------------------- 6
 Question # 16 (Multiple Choice)------------------------------------------------------------------------------------------- 7
Question # 17 (Multiple Choice)------------------------------------------------------------------------------------------- 7
Question # 18 (Multiple Choice)------------------------------------------------------------------------------------------- 7
Question # 19 (Sentence completion) ----------------------------------------------------------------------------------- 7
Question # 20 (Multiple Choice)------------------------------------------------------------------------------------------- 8
Questions and Answers ------------------------------------------------------------------------------------------------------ 9
Question # 1 (Multiple Choice) -------------------------------------------------------------------------------------------- 9
Question # 2 (Matching)----------------------------------------------------------------------------------------------------- 9
Question # 3 (Sentence completion) ------------------------------------------------------------------------------------ 10
Question # 4 (Sentence completion) ------------------------------------------------------------------------------------ 10
Question # 5 (Multiple Choice) ------------------------------------------------------------------------------------------- 11
Question # 6 (Matching)---------------------------------------------------------------------------------------------------- 11
Question # 7 (Sentence completion) ------------------------------------------------------------------------------------ 12
Question # 8 (Multiple Choice) ------------------------------------------------------------------------------------------- 12
Question # 9 (Sentence completion) ------------------------------------------------------------------------------------ 13
Question # 10 (Multiple Choice)------------------------------------------------------------------------------------------ 13
Question # 11 (Multiple Choice)------------------------------------------------------------------------------------------ 14
Question # 12 (Sentence completion) ---------------------------------------------------------------------------------- 14
Question # 13 (Multiple Choice)------------------------------------------------------------------------------------------ 15
Question # 14 (Multiple Choice)------------------------------------------------------------------------------------------ 15
Question # 15 (Matching) -------------------------------------------------------------------------------------------------- 16
Question # 16 (Multiple Choice)------------------------------------------------------------------------------------------ 16
Question # 17 (Multiple Choice)------------------------------------------------------------------------------------------ 17
Question # 18 (Multiple Choice)------------------------------------------------------------------------------------------ 17
Question # 19 (Sentence completion) ---------------------------------------------------------------------------------- 18
Question # 20 (Multiple Choice)------------------------------------------------------------------------------------------ 18
Questions

Question # 1 (Multiple Choice)

Which type of artificial intelligence workload uses sensors to proactively alert users about
potential equipment mechanical failures?

A. Anomaly detection
B. Computer vision
C. Natural language processing
D. Conversational AI

Question # 2 (Matching)
Match the technologies on the left to the correct descriptions on the right.

Technologies Descriptions
A. Object detection _____ 1. Identify the location of a moving car within an
B. Image classification image.
C. Optical character _____ 2. Detect and read car registration plates in an image.
recognition _____ 3. Differentiate types of vehicles from an image set
containing different vehicle types.

Question # 3 (Sentence completion)

Select the answer that correctly completes the sentence.

The principle that describes raising awareness of the limitations of responsible AI-based
solutions is called: ________________.

A. Privacy and security

B. Reliability and safety
C. Transparency
D. Accountability

Question # 4 (Sentence completion)

Select the answer that correctly completes the sentence.

The principle of providing the benefits of responsible AI systems to all parts of society regardless
of their gender or ethnicity is called: ____________.

A. Privacy and security

B. Reliability and safety
C. Inclusiveness
 D. Accountability

Question # 5 (Multiple Choice)

You need to identify numerical values that represent the probability of dogs becoming ill based
on their age and body fat percentage.

Which type of machine learning model should you use?

A. Linear regression
B. Multiple linear regression
C. Logistic regression
D. Hierarchical clustering

Question # 6 (Matching)
Match the machine learning algorithms on the left to the correct descriptions on the right.

Machine learning Descriptions

algorithms
A. Clustering _____ 1. Predict a numeric label based on an item’s features.
B. Regression _____ 2. Group similar items based on their features.
C. Classification _____ 3. Assign items into a set of predefined categories.

Question # 7 (Sentence completion)

Select the answer that correctly completes the sentence.

You plan to use machine learning to predict how ill dogs become based on their age and body
fat percentage.

The model should include ___________.

A. two features and one label

B. two labels and one feature
C. three labels
D. three features

Question # 8 (Multiple Choice)

You create a multiclass classification model.

You need to evaluate the model.

Which two evaluation metrics can you use? Each correct answer presents a complete solution.
 A. F1 score
B. Accuracy
C. Rand index
D. Mean Squared Error (MSE)

Question # 9 (Sentence completion)

Select the answer that correctly completes the sentence.

You train an Azure Machine Learning model and plan to deploy the model as a predictive
service in a production environment.

You must create an inference cluster before you deploy the model to _______________.

A. Azure Kubernetes Service

B. Azure Container Instance
C. an Azure Function
D. an Azure Logic Apps app

Question # 10 (Multiple Choice)

You plan to build and deploy a predictive model by using AutoML UI in Azure Machine Learning.

Which three machine learning tasks are supported? Each correct answer presents a complete
solution.

A. Classification
B. Regression
C. Forecasting
D. Clustering
E. Reinforcement learning

Question # 11 (Multiple Choice)

Which technique serves as the basis for modern image classification solutions?

A. Deep learning
B. Anomaly detection
C. Linear regression
D. Multiple linear regression
Question # 12 (Sentence completion)
Select the answer that correctly completes the sentence.

You can extract information printed on food product labels by using _____________________.

A. image classification
B. natural language processing
C. optical character recognition
D. image segmentation

Question # 13 (Multiple Choice)

What are two capabilities of the Azure Computer Vision service? Each correct answer presents a
complete solution.

A. Class prediction
B. Model training
C. Model authoring
D. Data visualization
E. Data exploration

Question # 14 (Multiple Choice)

You need to train a machine learning model to detect company logos in images.

What should you use?

A. Azure Custom Vision image classification

B. Azure Custom Vision object detection
C. Azure Face Service
D. Language Understanding Intelligent Service (LUIS)

Question # 15 (Matching)
Match the features on the left to the correct descriptions on the right.

Features Descriptions
A. Entity linking _____ 1. Evaluate the main points from the text in a
B. Sentiment analysis document.
C. Key phrase extraction _____ 2. Determine whether the content of a document is
D. Named entity positive or negative.
recognition _____ 3. Identify words in documents that represent persons,
locations, or organizations.
Question # 16 (Multiple Choice)
You need to collect the names of people, organizations, and events from a set of PDF
documents.

Which natural language processing feature should you use?

A. Extractive summarization
B. Sentiment analysis
C. Named entity recognition
D. Key phrase extraction

Question # 17 (Multiple Choice)

Which three capabilities does Azure Text Analytics service support? Each correct answer presents
a complete solution.

A. Unlimited document size

B. All world languages
C. Chatbot integration
D. Multilingual content
E. Confidence scoring

Question # 18 (Multiple Choice)

You need to identify users based on their voice.

Which Azure Speech service feature should you use?

A. Conversation transcription
B. Pronunciation assessment
C. Language Understanding Intelligent Service (LUIS)
D. Speaker recognition

Question # 19 (Sentence completion)

Select the answer that correctly completes the sentence.

You can exchange chatbot activities with other services by implementing ________________.

A. cards
B. channels
C. dialog
D. turns
Question # 20 (Multiple Choice)
You develop a chatbot by using QnA Maker.

You need to add a personality to the chatbot.

What should you do?

A. Provide a default answer.

B. Add chit-chat to the knowledge base.
C. Increase the Cognitive Search resource pricing tier limit.
D. Add hero cards to the chatbot.
Questions and Answers

Question # 1 (Multiple Choice)

Which type of artificial intelligence workload uses sensors to proactively alert users about
potential equipment mechanical failures?

A. Anomaly detection
B. Computer vision
C. Natural language processing
D. Conversational AI

Answer: A. Anomaly detection

Objective: 1.1 Identify features of common AI workloads
Rationale: Anomaly detection analyzes data collected over time to identify unusual
changes, such as fluctuations in engine revolutions or brake temperature.
Computer vision, natural language processing, or conversational AI do not
play a role in the scenario referenced in the question.
URL: https://docs.microsoft.com/en-us/learn/modules/get-started-ai-
fundamentals/1-introduction
https://docs.microsoft.com/en-us/learn/modules/get-started-ai-
fundamentals/3-understand-anomaly-detection

Question # 2 (Matching)
Match the technologies on the left to the correct descriptions on the right.

Technologies Descriptions
A. Object detection _____ 1. Identify the location of a moving car within an
B. Image classification image.
C. Optical character _____ 2. Detect and read car registration plates in an image.
recognition _____ 3. Differentiate types of vehicles from an image set
containing different vehicle types.
Answer: Object detection (A) matches description 1: Identify the
location of a moving car within an image.
Image classification (B) matches description 3: Differentiate
types of vehicles from an image set containing different
vehicle types.
Optical character recognition (C) matches description 2:
Detect and read car registration plates in an image.
Objective: 1.1 Identify features of common AI workloads
Rationale: Object detection allows you to identify specific object types
in images. Optical character recognition (OCR) allows you to
 identify text in images. Image classification allows you to
differentiate between different types of objects in images.
URL: https://docs.microsoft.com/en-us/learn/modules/get-
started-ai-fundamentals/4-understand-computer-vision

Question # 3 (Sentence completion)

Select the answer that correctly completes the sentence.

The principle that describes raising awareness of the limitations of responsible AI-based
solutions is called: ________________.

A. Privacy and security

B. Reliability and safety
C. Transparency
D. Accountability

Answer: C. Transparency
Objective: 1.2 Identify guiding principles for responsible AI
Rationale: Transparency provides clarity regarding the purpose of AI solutions, the way
they work, as well as their limitations. Other principles of responsible AI are
meant to apply to any AI solution, regardless of their limitations.
URL: Identify principles and practices for responsible AI - Learn | Microsoft Docs

Question # 4 (Sentence completion)

Select the answer that correctly completes the sentence.

The principle of providing the benefits of responsible AI systems to all parts of society regardless
of their gender or ethnicity is called: ____________.

A. Privacy and security

B. Reliability and safety
C. Inclusiveness
D. Accountability

Answer: C. Inclusiveness
Objective: 1.2 Identify guiding principles for responsible AI
Rationale: Responsible AI systems should empower everyone and engage people. AI
should bring benefits to all parts of society, regardless of physical ability,
gender, sexual orientation, ethnicity, or other factors ensuring inclusiveness.
URL: https://docs.microsoft.com/en-us/learn/modules/get-started-ai-
fundamentals/8-understand-responsible-ai
Question # 5 (Multiple Choice)
You need to identify numerical values that represent the probability of dogs becoming ill based
on their age and body fat percentage.

Which type of machine learning model should you use?

A. Linear regression
B. Multiple linear regression
C. Logistic regression
D. Hierarchical clustering

Answer: B. Multiple linear regression

Objective: 2.1 Identify common machine learning types
Rationale: Modeling relationships between several features and a single label is the
primary characteristic of multiple linear regression, in contrast with linear
regression which uses a single feature. Logistic regression is a classification
model and hierarchical clustering is a type of clustering algorithm.
URL: https://docs.microsoft.com/en-us/learn/modules/understand-regression-
machine-learning/4-multiple-linear-regression
https://docs.microsoft.com/en-us/learn/modules/understand-classification-
machine-learning/2-what-is-classification
https://docs.microsoft.com/en-us/learn/modules/train-evaluate-cluster-
models/4-different-types-clustering

Question # 6 (Matching)
Match the machine learning algorithms on the left to the correct descriptions on the right.

Machine learning Descriptions

algorithms
A. Clustering _____ 1. Predict a numeric label based on an item’s features.
B. Regression _____ 2. Group similar items based on their features.
C. Classification _____ 3. Assign items into a set of predefined categories.
Answer: Regression (B) matches description 1: Predict a numeric
label based on an item’s features.
Clustering (A) matches description 2: Group similar items
based on their features.
Classification (C) matches description 3: Assign items into a
set of predefined categories.
Objective: 2.1 Identify common machine learning types
Rationale: The regression technique is used to predict numeric values.
Classification predicts the category in which an input value
should be categorized. Clustering groups data points that
have similar characteristics.
 URL: https://docs.microsoft.com/en-us/learn/modules/create-
regression-model-azure-machine-learning-
designer/introduction
https://docs.microsoft.com/en-
us/learn/modules/understand-classification-machine-
learning/2-what-is-classification
https://docs.microsoft.com/en-us/learn/modules/create-
clustering-model-azure-machine-learning-
designer/introduction

Question # 7 (Sentence completion)

Select the answer that correctly completes the sentence.

You plan to use machine learning to predict how ill dogs become based on their age and body
fat percentage.

The model should include ___________.

A. two features and one label

B. two labels and one feature
C. three labels
D. three features

Answer: A. Two features and one label

Objective: 2.2 Describe core machine learning concepts
Rationale: The current scenario represents a model in which you are trying to establish a
relationship between two features (a dog's age and body fat percentage) and
one label (the likelihood of that dog becoming ill).
URL: https://docs.microsoft.com/en-us/learn/modules/understand-regression-
machine-learning/4-multiple-linear-regression

Question # 8 (Multiple Choice)

You create a multiclass classification model.

You need to evaluate the model.

Which two evaluation metrics can you use? Each correct answer presents a complete solution.

A. F1 score
B. Accuracy
C. Rand index
D. Mean Squared Error (MSE)

Answer: A. F1 score AND

B. Accuracy
 Objective: 2.2 Describe core machine learning concepts
Rationale: You can use the F1 score and accuracy metrics for evaluating classification
models. The F1 score combines precision and recall for classification
evaluation while accuracy evaluates the ratio of correct predictions. Rand
index is used for evaluating clustering models. MSE is used for evaluating
regression models.
URL: https://docs.microsoft.com/en-us/learn/modules/create-classification-model-
azure-machine-learning-designer/evaluate-model

Question # 9 (Sentence completion)

Select the answer that correctly completes the sentence.

You train an Azure Machine Learning model and plan to deploy the model as a predictive
service in a production environment.

You must create an inference cluster before you deploy the model to _______________.

A. Azure Kubernetes Service

B. Azure Container Instance
C. an Azure Function
D. an Azure Logic Apps app

Answer: A. Azure Kubernetes Service

Objective: 2.3 Identify core tasks in creating a machine learning solution

Rationale: In Azure Machine Learning, you have the option of deploying a predictive
service to Azure Container Instance (ACI) or Azure Kubernetes Service (AKS).
For production scenarios, you should use an AKS deployment, which requires
creating an inference cluster compute target. ACI-based deployment is
suitable for testing. Azure Machine Learning does not support deployment of
predictive services to Azure Functions or Azure Logic Apps.
URL: https://docs.microsoft.com/en-us/learn/modules/use-automated-machine-
learning/deploy-model

Question # 10 (Multiple Choice)

You plan to build and deploy a predictive model by using AutoML UI in Azure Machine Learning.

Which three machine learning tasks are supported? Each correct answer presents a complete
solution.

A. Classification
B. Regression
 C. Forecasting
D. Clustering
E. Reinforcement learning

Answer: A. Classification AND

B. Regression AND
C. Forecasting
Objective: 2.4 Describe capabilities of No-Code Machine Learning with Azure Machine
Learning studio
Rationale: AutoML UI supports classification, regression and forecasting machine
learning tasks. Clustering and reinforcement learning are not available on
AutoML UI.
URL: https://docs.microsoft.com/en-us/learn/modules/use-automated-machine-
learning/deploy-model

Question # 11 (Multiple Choice)

Which technique serves as the basis for modern image classification solutions?

A. Deep learning
B. Anomaly detection
C. Linear regression
D. Multiple linear regression

Answer: A. Deep learning

Objective: 3.1 Identify common types of computer vision solutions
Rationale: Modern image classification solutions are based on deep learning techniques
that make use of convolutional neural networks (CNNs) to identify patterns in
the pixels that comprise an image and map it to a particular class. Anomaly
detection is an Artificial Intelligence technique that detects unusual
occurrences in data patterns. Both linear and multiple linear regression are
regression techniques, rather than classifications.
URL: https://docs.microsoft.com/en-us/learn/modules/classify-images-custom-
vision/1a-overview-classification

Question # 12 (Sentence completion)

Select the answer that correctly completes the sentence.

You can extract information printed on food product labels by using _____________________.

A. image classification
B. natural language processing
C. optical character recognition
 D. image segmentation

Answer: C. Optical character recognition

Objective: 3.1 Identify common types of computer vision solution

Rationale: OCR supports printed text extraction. Image classification is a distractor

because it cannot detect text on an object. Natural language processing is the
approach of handling linguistic data, and it is a distractor. Image
segmentation is not correct since it is used to segment areas in an image.
URL: https://docs.microsoft.com/en-us/azure/cognitive-services/computer-
vision/overview-ocr

Question # 13 (Multiple Choice)

What are two capabilities of the Azure Computer Vision service? Each correct answer presents a
complete solution.

A. Class prediction
B. Model training
C. Model authoring
D. Data visualization
E. Data exploration

Answer: A. Class prediction AND

B. Model training
Objective: 3.2 Identify Azure tools and services for computer vision tasks
Rationale: Azure Computer Vision is an Azure resource that offers training and
prediction capabilities. It does not provide the ability to author models, nor
does it allow you to explore and visualize datasets.
URL: https://docs.microsoft.com/en-us/learn/modules/classify-images-custom-
vision/2-azure-image-classification

Question # 14 (Multiple Choice)

You need to train a machine learning model to detect company logos in images.

What should you use?

A. Azure Custom Vision image classification

B. Azure Custom Vision object detection
C. Azure Face Service
D. Language Understanding Intelligent Service (LUIS)

Answer: A. Azure Custom Vision object detection

Objective: 3.2 Identify Azure tools and services for computer vision tasks
 Rationale: Object detection functionality in Azure Custom Vision can identify logos in
images.
Image classification functionality in Azure Custom Vision is used for classifying
a set of images to groups.
Azure Face Service is specifically used for identifying faces. This service cannot
identify logos.
LUIS is used for understanding natural language.
URL: Detect objects in images with the Custom Vision service - Learn | Microsoft
Docs

Question # 15 (Matching)
Match the features on the left to the correct descriptions on the right.

Features Descriptions
A. Entity linking _____ 1. Evaluate the main points from the text in a
B. Sentiment analysis document.
C. Key phrase extraction _____ 2. Determine whether the content of a document is
D. Named entity positive or negative.
recognition _____ 3. Identify words in documents that represent persons,
locations, or organizations.
Answer: Key phrase extraction (C) matches description 1: Evaluate
the main points from the text in a document.
Sentiment analysis (B) matches description 2: Determine
whether the content of a document is positive or negative.
Named entity recognition (D) matches description 3: Identify
words in documents that represent persons, locations, or
organizations.
Objective: 4.1 Identify features of common NLP Workload Scenarios
Rationale: Key phrase extraction scans documents and identifies the
main points from the documents. Sentiment analysis scans a
document to determine whether the content is positive or
negative. Named entity recognition scans documents and
identified important entities (objects, nouns) in the
document.
URL: https://docs.microsoft.com/en-us/learn/modules/extract-
insights-text-with-text-analytics-service/

Question # 16 (Multiple Choice)

You need to collect the names of people, organizations, and events from a set of PDF
documents.

Which natural language processing feature should you use?

 A. Extractive summarization
B. Sentiment analysis
C. Named entity recognition
D. Key phrase extraction

Answer: C. Named entity recognition

Objective: 4.1 Identify features of common NLP Workload Scenarios

Rationale: The named entity recognition feature in text analytics identifies a range of
prebuilt entities such as people, places, and organizations which can be used
for entity identification in PDF documents.
URL: https://docs.microsoft.com/en-us/learn/modules/get-started-ai-
fundamentals/5-understand-natural-language-process

Question # 17 (Multiple Choice)

Which three capabilities does Azure Text Analytics service support? Each correct answer presents
a complete solution.

A. Unlimited document size

B. All world languages
C. Chatbot integration
D. Multilingual content
E. Confidence scoring

Answer: C. Chatbot integration AND

D. Multilingual content AND
E. Confidence scoring
Objective: 4.2 Identify Azure tools and services for NLP workloads
Rationale: Azure Text Analytics supports chatbot integration, multilingual content, and
confidence scoring. It recognizes about 120 languages. Document sizes must
be under 5,120 characters.
URL: https://docs.microsoft.com/en-us/learn/modules/extract-insights-text-with-
text-analytics-service/3-detect-language

Question # 18 (Multiple Choice)

You need to identify users based on their voice.

Which Azure Speech service feature should you use?

A. Conversation transcription
B. Pronunciation assessment
C. Language Understanding Intelligent Service (LUIS)
 D. Speaker recognition

Answer: D. Speaker recognition

Objective: 4.2 Identify Azure tools and services for NLP workloads

Rationale: The Speaker Recognition service provides algorithms that verify and identify
speakers by their unique voice characteristics.
The Conversation transcription service is used for generating transcripts of a
conversation.
Pronunciation assessment evaluates the accuracy and fluency of spoken
audio.
LUIS is a distractor because it is used for understanding natural language
which cannot be used for differentiating speakers
URL: Recognize and synthesize speech - Learn | Microsoft Docs

Question # 19 (Sentence completion)

Select the answer that correctly completes the sentence.

You can exchange chatbot activities with other services by implementing ________________.

A. cards
B. channels
C. dialog
D. turns

Answer: A. Channels
Objective: Exam objective number and text
Rationale: Activities are exchanged across channels, such as web chat, email, or
Microsoft Teams. Cards are visual elements used to contain messages. Dialog
is formed by a flow of activities. Activities are performed in turns, leading to a
user interaction with a chatbot.
URL: https://docs.microsoft.com/en-us/learn/modules/design-bot-conversation-
flow/1-introduction

Question # 20 (Multiple Choice)

You develop a chatbot by using QnA Maker.

You need to add a personality to the chatbot.

What should you do?

A. Provide a default answer.

B. Add chit-chat to the knowledge base.
C. Increase the Cognitive Search resource pricing tier limit.
D. Add hero cards to the chatbot.

Answer: B. Add chit-chat to the knowledge base.

Objective: 5.2 Identify Azure services for conversational AI

Rationale: You can add personality to a chatbot by providing answers that use a specific
conversational tone. You use the chit-chat feature to add the answers to a
chatbot knowledge base.
Provide a default answer from settings is incorrect because it is used to
provide a pre-set answer from a chatbot.
Increasing Cognitive Search resource pricing tier limits only increases the
number of concurrent requests that can connect to a chatbot.
Hero cards are used for showing media inside a chatbot, not to add a
personality.
URL: Build a QnA solution with QnA Maker - Learn | Microsoft Docs