View Full Microsoft AI-901 Exam Dumps and Practice Test Dumps

 

Question 101. Which Azure AI service can analyze the visual content of an image?

1. Azure AI Vision
2. Azure AI Speech
3. Azure AI Translator
4. Azure AI Language

Correct Answer: 1. Azure AI Vision

Explanation:

Azure AI Vision provides capabilities for analyzing and understanding visual information contained in images. Applications can use it to identify objects, generate image descriptions, detect visual features, extract text, and perform other image-related analysis tasks. For example, an application could analyze a photograph and identify objects or recognize printed text appearing in the image. Azure AI Speech is intended for audio and speech processing, while Azure AI Translator focuses on translating text between languages. Azure AI Language provides capabilities for understanding written language. Therefore, when a scenario requires an Azure service to interpret or analyze image content, Azure AI Vision is the appropriate choice.

Question 102. Which feature can determine the dominant colors and other visual properties of an image?

1. Speech recognition
2. Image analysis
3. Sentiment analysis
4. Entity recognition

Correct Answer: 2. Image analysis

Explanation:

Image analysis examines the visual characteristics and content of an image and can provide information about different visual properties. Depending on the capability being used, image analysis can identify objects, generate tags, describe scenes, detect visual characteristics, and identify features such as dominant colors. These capabilities allow applications to process images automatically rather than requiring a person to manually inspect every image. Speech recognition is used for converting spoken language into text, sentiment analysis evaluates opinions in written content, and entity recognition identifies named entities. Therefore, when a requirement involves understanding visual properties or content within an image, image analysis is the relevant AI capability.

Question 103. Which Azure AI Language feature extracts important words and phrases from a document?

1. Sentiment analysis
2. Named entity recognition
3. Key phrase extraction
4. Language detection

Correct Answer: 3. Key phrase extraction

Explanation:

Key phrase extraction identifies the main concepts or important phrases within a body of text. It can help applications quickly determine the subjects discussed in documents, reviews, emails, or other written content. For example, if a customer provides a long product review, key phrase extraction might identify phrases related to product quality, delivery, pricing, or customer service. Named entity recognition serves a different purpose by identifying specific entities such as people, organizations, and locations. Sentiment analysis evaluates the tone or opinion expressed in text, while language detection determines the language used. Therefore, key phrase extraction is the appropriate Azure AI Language capability when the requirement is to identify important terms or concepts.

Question 104. An application receives text in unknown languages. Which capability can identify the language used?

1. OCR
2. Text-to-speech
3. Key phrase extraction
4. Language detection

Correct Answer: 4. Language detection

Explanation:

Language detection is used to determine which human language is represented in a piece of text. This capability is useful when an application receives multilingual content without knowing the language in advance. After identifying the language, the application can choose an appropriate processing workflow, such as sending the text to translation, sentiment analysis, or another language-processing capability. OCR is intended to extract text from images, while text-to-speech converts written text into spoken audio. Key phrase extraction identifies important phrases but does not primarily determine the language. Therefore, if an application needs to automatically identify whether incoming text is English, Spanish, French, or another supported language, language detection is the appropriate capability.

Question 105. Which Azure AI Speech capability converts written text into spoken audio?

1. Text-to-speech
2. Speech-to-text
3. OCR
4. Object detection

Correct Answer: 1. Text-to-speech

Explanation:

Text-to-speech converts written text into synthesized spoken audio. Azure AI Speech can generate speech from application-provided text, making it useful for virtual assistants, accessibility solutions, educational applications, automated announcements, and other scenarios where users need to hear information rather than read it. Speech-to-text performs the opposite operation by converting spoken language into written text. OCR extracts text from visual documents or images, while object detection identifies objects within images. Understanding the direction of the conversion is important for AI-901 questions. If the input is written text and the desired output is spoken audio, text-to-speech is the correct Azure AI Speech capability.

Question 106. Which Azure AI Speech feature is useful for transcribing a recorded meeting?

1. Text-to-speech
2. Speech-to-text
3. Image analysis
4. Key phrase extraction

Correct Answer: 2. Speech-to-text

Explanation:

Speech-to-text can convert spoken words from an audio recording into written text, making it useful for meeting transcription and other voice-processing scenarios. A recorded meeting can contain multiple spoken statements that need to be transformed into text so they can be searched, reviewed, stored, or processed further. Azure AI Speech provides speech recognition capabilities for these types of applications. Text-to-speech would instead create spoken audio from written text. Image analysis processes visual information, while key phrase extraction analyzes written content after text has been obtained. Therefore, when the primary requirement is to create a written transcript from recorded speech, speech-to-text is the appropriate capability.

Question 107. Which service should an application use to translate a user’s message into another supported language?

1. Azure AI Vision
2. Azure AI Search
3. Azure AI Translator
4. Azure Machine Learning

Correct Answer: 3. Azure AI Translator

Explanation:

Azure AI Translator is designed to perform machine translation between supported human languages. An application can send text to the translation service and receive translated content that can be presented to users or used in another processing workflow. This capability is useful for multilingual websites, customer-support applications, communication tools, and international business systems. Azure AI Vision focuses on visual analysis, Azure AI Search provides search and retrieval capabilities, and Azure Machine Learning supports machine learning development and deployment. Therefore, when a scenario explicitly requires converting written content from one language into another, Azure AI Translator is the Azure service that directly addresses the requirement.

Question 108. Which Azure AI service can extract fields such as invoice number, date, and total amount from invoices?

1. Azure AI Speech
2. Azure AI Vision
3. Azure AI Search
4. Azure AI Document Intelligence

Correct Answer: 4. Azure AI Document Intelligence

Explanation:

Azure AI Document Intelligence is designed to process documents and extract useful information from structured and semi-structured content. Invoice processing is a common example because invoices can contain fields such as invoice numbers, dates, vendor details, line items, taxes, and totals. Instead of requiring an employee to manually enter these values, a document-processing solution can analyze the document and return structured information for further processing. Azure AI Vision can provide general image and OCR capabilities, but Document Intelligence is specifically suited to document understanding and field extraction. Azure AI Search focuses on information retrieval. Therefore, invoice field extraction is an appropriate use case for Azure AI Document Intelligence.

Question 109. What is the purpose of indexing documents in Azure AI Search?

1. To make information searchable and retrievable
2. To convert documents into speech
3. To train a classification model automatically
4. To detect faces in images

Correct Answer: 1. To make information searchable and retrievable

Explanation:

Indexing organizes information so that Azure AI Search can efficiently search and retrieve relevant content. When documents and their searchable fields are indexed, applications can query the index to locate information matching a user’s search request. This is especially useful for enterprise search applications and generative AI solutions that need to retrieve relevant organizational information. Indexing itself is not the same as model training, speech synthesis, or face detection. In retrieval-based AI applications, indexed content can provide useful context for downstream processing. Therefore, when an AI-901 scenario asks why documents are indexed in Azure AI Search, the main purpose is to make their information searchable and retrievable.

Question 110. Which approach can help a generative AI application answer questions using an organization’s internal documents?

1. Increase image resolution
2. Use retrieval and grounding with relevant documents
3. Convert all documents into audio
4. Disable search capabilities

Correct Answer: 2. Use retrieval and grounding with relevant documents

Explanation:

Retrieval and grounding can help a generative AI application use information contained in an organization’s internal documents when generating responses. A typical workflow can retrieve relevant passages from a search index or another data source and provide those passages as context to the language model. This helps the model generate an answer based on the organization’s information instead of relying exclusively on its general training knowledge. The approach is particularly useful for enterprise question-answering systems, internal knowledge assistants, and document-based applications. Image resolution and audio conversion do not directly provide textual context to the language model. Therefore, retrieval combined with grounding is an appropriate approach.

Question 111. Which Azure Machine Learning feature can help compare different machine learning models during experimentation?

1. Model evaluation and experiment tracking
2. Text-to-speech
3. OCR
4. Language translation

Correct Answer: 1. Model evaluation and experiment tracking

Explanation:

Model evaluation and experiment tracking help data scientists compare machine learning experiments and understand how different models perform. During development, multiple algorithms, configurations, or feature sets may be tested. Tracking the results allows developers to examine evaluation metrics and identify which configurations meet the requirements of the workload. This process is different from language translation, OCR, and text-to-speech, which are capabilities provided by specialized Azure AI services. Azure Machine Learning provides tools and workflows that support experimentation and model development. For AI-901, when the scenario involves comparing model performance across experiments, evaluation and experiment tracking are important concepts to consider.

Question 112. What is a hyperparameter in machine learning?

1. A value automatically generated only after deployment
2. A setting chosen to control the model training process
3. A label assigned to every prediction
4. A document stored in Azure AI Search

Correct Answer: 2. A setting chosen to control the model training process

Explanation:

A hyperparameter is a setting that controls aspects of the machine learning training process and is generally selected before or during training rather than learned directly as a model parameter from the training examples. Examples can include learning rate, number of trees in some ensemble models, maximum tree depth, or the number of clusters in certain algorithms. Choosing suitable hyperparameter values can affect model performance and training behavior. Hyperparameters should not be confused with labels, which represent target values in supervised learning. They are also unrelated to documents in Azure AI Search. Understanding the difference between model parameters and hyperparameters is an important machine learning concept for AI-901.

Question 113. Which machine learning problem predicts a category such as “fraud” or “not fraud”?

1. Regression
2. Clustering
3. Classification
4. Dimensionality reduction

Correct Answer: 3. Classification

Explanation:

Classification is a supervised machine learning task in which a model predicts a discrete category or class. For example, a fraud-detection model may classify transactions as fraudulent or legitimate. Other examples include identifying whether an email is spam, determining whether an image belongs to a particular category, or predicting whether a customer will respond to an offer. Regression differs because it predicts numerical values such as prices or temperatures. Clustering groups similar examples without predefined labels, while dimensionality reduction transforms data into fewer dimensions while attempting to preserve useful information. Therefore, when the prediction consists of categories such as “fraud” and “not fraud,” classification is the appropriate machine learning problem.

Question 114. Which machine learning task predicts a continuous numerical value such as house price?

1. Classification
2. Regression
3. Clustering
4. Entity recognition

Correct Answer: 2. Regression

Explanation:

Regression is a supervised machine learning task used to predict continuous numerical values. A house-price prediction model is a common example because the output may be a numerical amount such as a property’s estimated market value. Other examples include predicting sales revenue, temperature, demand, or delivery time. Classification instead predicts discrete categories, such as whether a transaction is fraudulent. Clustering groups similar observations without requiring predefined target labels. Entity recognition is a natural language processing capability rather than a general numerical prediction task. For AI-901, if the target value is numerical and can vary continuously, regression is the appropriate machine learning approach.

Question 115. Which machine learning technique groups similar data points without predefined labels?

1. Clustering
2. Regression
3. Classification
4. Speech recognition

Correct Answer: 1. Clustering

Explanation:

Clustering is an unsupervised machine learning technique that groups data points based on similarities in their characteristics. Unlike supervised learning, clustering does not require predefined target labels for the examples. It can be useful for customer segmentation, grouping documents, identifying patterns in data, and discovering natural structures within a dataset. Classification requires labeled examples to predict predefined categories, while regression predicts numerical values. Speech recognition is an AI capability for processing spoken language rather than a general clustering technique. Therefore, when a scenario describes grouping similar records without providing predefined categories, clustering is the appropriate machine learning approach.

Question 116. Which metric is particularly useful when both false positives and false negatives need to be considered in a classification model?

1. Mean absolute error
2. F1 score
3. Root mean squared error
4. Mean squared error

Correct Answer: 2. F1 score

Explanation:

The F1 score combines precision and recall into a single metric and is useful when both false positives and false negatives are important. Precision focuses on the correctness of positive predictions, while recall focuses on how many actual positive cases are successfully identified. The F1 score represents the harmonic mean of precision and recall, providing a balanced measure when both aspects matter. Mean absolute error, mean squared error, and root mean squared error are commonly associated with regression evaluation. Therefore, when an AI-901 classification scenario requires a metric that balances precision and recall, the F1 score is an appropriate choice.

Question 117. What does recall measure in a binary classification model?

1. The percentage of all predictions that are correct
2. The average numerical prediction error
3. The proportion of actual positive cases correctly identified
4. The percentage of predicted positives that are correct

Correct Answer: 3. The proportion of actual positive cases correctly identified

Explanation:

Recall measures the proportion of actual positive cases that a classification model successfully identifies. Its formula is true positives divided by the sum of true positives and false negatives. Recall is particularly important in situations where missing a positive case can have significant consequences. For example, a screening system may prioritize identifying as many relevant positive cases as possible. Precision has a different focus because it measures the proportion of predicted positive cases that are actually positive. Accuracy measures overall correctness across classes. Therefore, when the requirement is to determine how many actual positive cases were successfully detected, recall is the appropriate metric.

Question 118. Which Azure AI principle emphasizes making people responsible for the decisions and actions of an AI system?

1. Inclusiveness
2. Transparency
3. Fairness
4. Accountability

Correct Answer: 4. Accountability

Explanation:

Accountability means that people and organizations remain responsible for the development, deployment, operation, and outcomes of AI systems. AI systems may automate decisions or recommendations, but organizations still need appropriate processes for monitoring performance, addressing issues, and determining who is responsible for the system’s behavior. Transparency is related to making AI systems and their decisions understandable, while fairness focuses on avoiding unjustified differences in outcomes. Inclusiveness focuses on designing systems that can serve people with different abilities and circumstances. Therefore, when an AI-901 question asks which responsible AI principle emphasizes responsibility for AI systems and their outcomes, accountability is the appropriate answer.

Question 119. Which practice can improve transparency in an AI solution?

1. Document the system’s purpose, limitations, and decision process
2. Hide all information about the model
3. Remove evaluation metrics
4. Prevent users from understanding system outputs

Correct Answer: 1. Document the system’s purpose, limitations, and decision process

Explanation:

Transparency involves making information about an AI system understandable and accessible to appropriate stakeholders. Documenting the system’s purpose, capabilities, limitations, data considerations, and relevant decision processes can help users and responsible teams understand how the system should be used and what its outputs mean. Transparency does not require revealing every technical implementation detail to every user, but important information should be communicated clearly enough to support informed use and oversight. Hiding information and removing evaluation metrics can make an AI system harder to understand and monitor. Therefore, documenting the purpose, limitations, and decision process is a practical way to support transparency.

Question 120. What is the primary purpose of deploying a trained machine learning model to an online endpoint?

1. To permanently delete the training data
2. To translate all model inputs
3. To provide predictions through a callable service interface
4. To label the original training dataset

Correct Answer: 3. To provide predictions through a callable service interface

Explanation:

An online endpoint allows a trained machine learning model to be made available for real-time inference requests. Applications can send input data to the endpoint and receive predictions or other model outputs in response. This makes the model usable by applications, websites, business processes, or other services that require predictions while operating. Deploying a model to an endpoint does not automatically delete training data, translate model inputs, or label the original dataset. Those are separate activities within a broader machine learning workflow. For AI-901, when a scenario asks how a trained model can serve predictions to applications through requests, an online endpoint is the relevant deployment concept.