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

 

Question 141. Which Azure AI service can identify objects and describe visual content in images?

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

Correct Answer: 3. Azure AI Vision

Explanation:

Azure AI Vision provides capabilities for analyzing visual information contained in images. It can identify objects, generate descriptions, assign tags, and extract other useful information from visual content. These capabilities allow applications to understand images without requiring every image to be manually reviewed by a person. For example, a business could analyze product photographs and automatically identify objects or visual characteristics. Azure AI Speech focuses on speech processing, Azure AI Language works with written language, and Azure AI Translator provides translation. Therefore, when a scenario requires an Azure service to analyze and understand visual content in images, Azure AI Vision is the appropriate choice.

Question 142. Which capability can identify the location of objects within an image?

1. Object detection
2. Sentiment analysis
3. Language detection
4. Speech synthesis

Correct Answer: 1. Object detection

Explanation:

Object detection identifies objects within an image and can determine where those objects are located. Unlike simple image classification, which assigns a category to an entire image, object detection can identify multiple objects and provide their approximate locations, commonly represented using bounding boxes. This makes object detection useful for scenarios such as identifying products on shelves, detecting vehicles in traffic images, or locating specific items in photographs. Sentiment analysis evaluates written opinions, language detection identifies languages, and speech synthesis converts text into audio. Therefore, when an AI-901 scenario requires identifying objects and their positions inside an image, object detection is the appropriate computer vision capability.

Question 143. Which Azure AI capability can divide an image into regions belonging to different categories?

1. OCR
2. Image segmentation
3. Speech recognition
4. Entity recognition

Correct Answer: 2. Image segmentation

Explanation:

Image segmentation separates an image into meaningful regions or pixels according to their characteristics or categories. Instead of simply identifying that an object exists somewhere in an image, segmentation can identify the specific pixels that belong to an object or region. This makes segmentation useful in scenarios where precise boundaries are important. For example, an application may need to distinguish road surfaces, vehicles, buildings, or other regions within an image. OCR extracts text, speech recognition processes spoken language, and entity recognition identifies named entities in text. Therefore, when a requirement involves assigning image pixels to different regions or categories, image segmentation is the appropriate capability.

Question 144. Which Azure AI Language capability can summarize the main ideas of a long document?

1. Entity recognition
2. Language detection
3. Key phrase extraction
4. Text summarization

Correct Answer: 4. Text summarization

Explanation:

Text summarization is designed to produce a shorter version of a document while preserving important information and main ideas. It can be useful when users need to understand lengthy reports, articles, support conversations, or other documents without reading every sentence. A summarization capability can help applications create concise versions of large amounts of text for easier review. Entity recognition identifies specific entities, language detection identifies the language, and key phrase extraction identifies important phrases rather than generating a complete summary. Therefore, when the requirement is to reduce a long document into a concise representation of its main points, text summarization is the appropriate language-processing capability.

Question 145. Which capability can identify whether text is written in English, French, or another supported language?

1. Language detection
2. Sentiment analysis
3. OCR
4. Text-to-speech

Correct Answer: 1. Language detection

Explanation:

Language detection determines the language used in a piece of text. This capability is especially useful for applications that receive multilingual content without knowing in advance which language each input uses. After identifying the language, an application can select an appropriate workflow, such as translation, sentiment analysis, summarization, or another language-processing task. Sentiment analysis evaluates opinions or emotional tone, OCR extracts text from images, and text-to-speech converts written text into spoken audio. Therefore, if a scenario asks an application to determine whether incoming text is English, French, German, or another supported language, language detection is the appropriate capability.

Question 146. Which feature allows an application to extract important concepts from a large amount of text?

1. Speech-to-text
2. Key phrase extraction
3. Object detection
4. Face detection

Correct Answer: 2. Key phrase extraction

Explanation:

Key phrase extraction identifies the most important concepts or phrases in a body of text. It can help applications quickly understand the main subjects discussed in documents, customer reviews, surveys, emails, or other written content. For example, a customer review might contain several sentences, but key phrase extraction could identify concepts related to product quality, shipping, pricing, or service. Speech-to-text converts spoken audio into written text, while object detection and face detection are computer vision capabilities. Therefore, when an AI-901 scenario asks for a capability that can identify significant concepts or phrases within text, key phrase extraction is the appropriate choice.

Question 147. Which Azure AI Speech capability can recognize spoken words and produce a transcript?

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 converts spoken language into written text. Azure AI Speech can process speech from supported audio sources and generate a transcript that applications can store, search, analyze, or display to users. This capability is useful for meeting transcription, voice commands, customer service applications, accessibility solutions, and other scenarios where spoken information needs to become machine-readable text. Text-to-speech performs the reverse process by converting written text into synthesized speech. Image analysis processes visual information, while key phrase extraction analyzes text after it has been obtained. Therefore, when the requirement is to recognize spoken words and create a written transcript, speech-to-text is the correct capability.

Question 148. Which service is appropriate for translating a product description into another language?

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

Correct Answer: 3. Azure AI Translator

Explanation:

Azure AI Translator is designed to translate text between supported human languages. A product description can be submitted to the translation service and converted into another language so that it can be presented to customers in different regions. This can support multilingual websites, e-commerce applications, customer communications, and international business processes. Azure AI Vision focuses on image understanding, Azure AI Speech handles speech and audio capabilities, and Azure Machine Learning provides tools for developing and managing machine learning models. Therefore, if the primary requirement is to translate written product information from one language to another, Azure AI Translator is the most directly relevant Azure service.

Question 149. Which Azure AI service can extract tables and fields from supported documents?

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

Correct Answer: 1. Azure AI Document Intelligence

Explanation:

Azure AI Document Intelligence is designed to analyze documents and extract structured information from their contents. Depending on the document and model being used, it can identify fields, tables, text, and other document elements. This makes it useful for processing invoices, receipts, forms, business records, and other documents where manually entering information would require significant effort. Azure AI Speech is focused on audio, Translator handles language translation, and Azure AI Search is primarily concerned with searching and retrieving indexed information. Therefore, when an AI-901 scenario requires extracting structured fields or tables from supported documents, Azure AI Document Intelligence is the appropriate service.

Question 150. Which Azure AI Search feature helps return information based on the meaning of a user’s query?

1. OCR
2. Semantic search
3. Speech synthesis
4. Object detection

Correct Answer: 2. Semantic search

Explanation:

Semantic search is designed to improve search results by considering the meaning and context of a query rather than relying only on exact keyword matching. This can help an application return information that is conceptually relevant even when the wording in the query differs from the wording in the indexed content. Semantic search can be useful in enterprise search and AI-powered knowledge applications where users ask natural-language questions. OCR extracts text from images, speech synthesis produces audio, and object detection identifies objects in images. Therefore, when an AI-901 scenario asks for search based on the meaning and context of a user’s query, semantic search is the relevant capability.

Question 151. What is the main benefit of using retrieval-augmented generation with enterprise documents?

1. It eliminates the need for any language model
2. It converts all documents into images
3. It provides relevant external information as context for generation
4. It guarantees every generated answer is correct

Correct Answer: 3. It provides relevant external information as context for generation

Explanation:

Retrieval-augmented generation, commonly called RAG, combines information retrieval with generative AI. A typical RAG solution first retrieves relevant content from an organization’s data source, such as an Azure AI Search index, and then provides that content to a generative model as contextual information. This can help the model generate responses that are based on the organization’s documents rather than relying exclusively on its general training knowledge. RAG does not eliminate the language model and does not guarantee that every generated response will be correct. Therefore, providing relevant external information as context is the main benefit described in this scenario.

Question 152. Which Azure Machine Learning feature provides a way to expose a trained model for real-time predictions?

1. Online endpoint
2. Data asset
3. Model registry
4. Training dataset

Correct Answer: 1. Online endpoint

Explanation:

An online endpoint allows a trained machine learning model to provide predictions in response to real-time requests. An application can send input data to the endpoint and receive the model’s prediction or output. This makes online endpoints useful for scenarios such as real-time fraud scoring, recommendation systems, classification services, and other applications that require immediate inference. A data asset manages a reference to data, while the model registry manages registered models and their versions. A training dataset is used to train the model rather than expose it for application requests. Therefore, an online endpoint is the appropriate Azure Machine Learning deployment capability for real-time predictions.

Question 153. What is the primary purpose of a batch endpoint in Azure Machine Learning?

1. To process large amounts of data asynchronously
2. To translate text instantly
3. To identify objects in a single photograph
4. To create speech from text

Correct Answer: 1. To process large amounts of data asynchronously

Explanation:

A batch endpoint is designed for processing larger amounts of input data asynchronously rather than responding immediately to individual real-time requests. This is useful when predictions can be generated as a batch job and immediate responses are not required. For example, an organization might use batch inference to score a large collection of records overnight or process a large set of images. Online endpoints are better suited to applications that need real-time predictions. Translation, image analysis, and speech synthesis are separate Azure AI capabilities. Therefore, when an AI-901 scenario describes processing a large dataset without requiring immediate responses for each request, a batch endpoint is an appropriate deployment approach.

Question 154. Which factor should be considered when selecting a machine learning evaluation metric?

1. Only the model’s file size
2. The business impact of different types of errors
3. The color of the user interface
4. The name of the dataset

Correct Answer: 2. The business impact of different types of errors

Explanation:

The choice of a machine learning evaluation metric should reflect the problem and the consequences of different prediction errors. For example, in some classification scenarios, false positives may be more costly than false negatives, while in other situations missing a positive case may be more serious. Metrics such as precision, recall, accuracy, and F1 score provide different perspectives on classification performance. Regression problems use metrics such as MAE, MSE, and RMSE. Therefore, developers should select metrics based on the model’s purpose and the impact of prediction errors rather than arbitrary factors such as interface color or dataset naming.

Question 155. Which metric is useful when large regression errors should receive greater penalty?

1. Accuracy
2. Precision
3. Mean squared error
4. Recall

Correct Answer: 3. Mean squared error

Explanation:

Mean squared error, or MSE, calculates the average of squared differences between predicted and actual numerical values. Because each error is squared, larger errors contribute disproportionately more to the final metric. This makes MSE useful when large prediction errors should receive greater emphasis during model evaluation. MAE also measures prediction error but uses absolute differences, so it does not amplify larger errors in the same way. Accuracy, precision, and recall are primarily classification metrics. Therefore, when a regression scenario specifically emphasizes giving greater penalty to larger prediction errors, mean squared error is an appropriate evaluation metric.

Question 156. Which Azure Machine Learning capability helps automate the selection of suitable machine learning models?

1. Automated ML
2. Azure AI Speech
3. Azure AI Translator
4. Azure AI Vision

Correct Answer: 1. Automated ML

Explanation:

Automated Machine Learning helps automate parts of the machine learning model development process. Depending on the task, Automated ML can try different algorithms, configurations, and preprocessing approaches and evaluate candidate models using selected metrics. This can help developers and data scientists identify suitable models without manually testing every possible combination. It does not mean that human oversight or evaluation is unnecessary, but it can reduce repetitive experimentation. Azure AI Speech, Translator, and Vision are specialized AI services with different purposes. Therefore, when a scenario asks for a capability that automates experimentation and model selection, Automated ML is the appropriate Azure Machine Learning feature.

Question 157. Which action can help prevent unauthorized access to an AI solution’s data?

1. Disable authentication
2. Apply appropriate identity and access controls
3. Publish all data publicly
4. Share credentials between users

Correct Answer: 2. Apply appropriate identity and access controls

Explanation:

Identity and access controls help ensure that only authorized users, applications, or services can access protected resources. In an AI solution, these controls can be applied to data stores, machine learning resources, endpoints, and other components. Organizations should follow appropriate access policies and grant users only the permissions required for their responsibilities. Disabling authentication, publishing sensitive data, or sharing credentials can increase security risks. Access control is therefore an important part of protecting data and AI resources. For AI-901, a scenario involving prevention of unauthorized access should point toward identity management, authentication, authorization, and least-privilege access practices.

Question 158. Which responsible AI principle focuses on making AI system behavior understandable to users?

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

Correct Answer: 1. Transparency

Explanation:

Transparency focuses on making information about an AI system understandable to appropriate users and stakeholders. This can include explaining the system’s purpose, capabilities, limitations, and relevant factors that influence its outputs. Clear documentation and communication can help users understand what an AI system can and cannot reliably do. Transparency does not necessarily mean exposing every technical implementation detail, but it supports informed use and appropriate oversight. Fairness addresses equitable outcomes, inclusiveness considers the needs of diverse users, and reliability concerns dependable operation. Therefore, when an AI-901 scenario focuses on helping users understand how an AI system behaves and what its limitations are, transparency is the appropriate responsible AI principle.

Question 159. What is the purpose of evaluating a machine learning model on unseen data?

1. To determine whether the model generalizes beyond its training examples
2. To guarantee that the model has no bias
3. To increase the number of training labels automatically
4. To eliminate the need for monitoring

Correct Answer: 1. To determine whether the model generalizes beyond its training examples

Explanation:

Evaluating a machine learning model on unseen data helps determine how well it generalizes beyond the examples used during training. A model may perform very well on training data because it has learned the training examples closely, but that does not necessarily mean it will perform equally well on new real-world data. Using validation and test datasets provides opportunities to evaluate model behavior on examples that were not used to fit the model. Unseen-data evaluation does not guarantee that a model is unbiased and does not eliminate the need for monitoring after deployment. Therefore, assessing generalization is the main reason for evaluating a model on unseen data.

Question 160. Which action is appropriate when a deployed model’s performance decreases because real-world data has changed?

1. Ignore the performance change
2. Delete all evaluation metrics
3. Investigate the change and consider retraining or updating the model
4. Remove all production monitoring

Correct Answer: 3. Investigate the change and consider retraining or updating the model

Explanation:

When a deployed model’s performance decreases because real-world data or relationships have changed, the organization should investigate the cause and determine whether the model needs to be updated or retrained. Changes in customer behavior, market conditions, operating environments, or input-data distributions can make a previously trained model less effective. Monitoring can help identify these changes, while evaluation can help determine their impact. Simply ignoring the problem or removing monitoring would make the situation harder to manage. Therefore, investigating the change and considering appropriate model updates or retraining is a sensible machine learning lifecycle response when production performance declines.