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

 

Question 41. Which Azure AI service can extract text from images and scanned documents?

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

Correct Answer: 1. Azure AI Vision

Explanation:

Azure AI Vision provides computer vision capabilities that can analyze visual content, including extracting printed or handwritten text from images through supported optical character recognition capabilities. This can be useful for processing scanned documents, photographs containing text, receipts, signs, and other visual material. The extracted text can then be passed to other services for additional processing, such as language analysis or search. Azure AI Speech is focused on spoken audio, Azure AI Language analyzes text, and Azure AI Search provides search and retrieval capabilities. Therefore, Azure AI Vision is the appropriate Azure AI service when the requirement is to extract text directly from images or scanned visual content.

Question 42. Which Azure AI Language capability detects the language used in a text document?

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

Correct Answer: 2. Language detection

Explanation:

Language detection identifies the language represented in a supplied text. This capability can be useful when an application receives content from users who may communicate in different languages and the application needs to determine which language-processing workflow should be used. For example, a customer-support application could detect whether a message is written in English, Spanish, French, or another supported language before applying additional processing. Entity recognition identifies important entities, key phrase extraction identifies significant terms, and sentiment analysis evaluates emotional polarity. Therefore, when the requirement is to determine which language a text is written in, the appropriate Azure AI Language capability is language detection.

Question 43. Which Azure AI service can convert text into synthesized speech?

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

Correct Answer: 3. Azure AI Speech

Explanation:

Azure AI Speech provides speech synthesis capabilities that can convert written text into generated spoken audio. Applications can use this capability for virtual assistants, accessibility features, automated announcements, educational applications, navigation systems, and other scenarios where information needs to be presented through speech. Speech services can support multiple languages and voices, depending on the available service configuration. Azure AI Search focuses on information retrieval, Azure AI Vision analyzes visual content, and Azure Machine Learning provides a broader platform for developing and deploying machine learning solutions. Therefore, Azure AI Speech is the appropriate service when an application needs to convert written text into synthesized speech.

Question 44. Which AI workload is used to determine whether a customer’s review is positive, neutral, or negative?

1. Sentiment analysis
2. Object detection
3. Image classification
4. Speech synthesis

Correct Answer: 1. Sentiment analysis

Explanation:

Sentiment analysis is a natural language processing workload that determines the sentiment expressed in a piece of text. Depending on the service and model, sentiment can be categorized into values such as positive, negative, or neutral, and additional confidence information may also be available. This capability is useful for analyzing customer reviews, feedback forms, support conversations, and survey responses at scale. Object detection and image classification are computer vision tasks, while speech synthesis converts text into spoken audio. Therefore, if an application needs to determine whether a customer’s written review expresses positive, neutral, or negative sentiment, sentiment analysis is the appropriate AI workload.

Question 45. Which computer vision task separates an image into meaningful regions or objects at the pixel level?

1. Classification
2. Object detection
3. Segmentation
4. Regression

Correct Answer: 3. Segmentation

Explanation:

Image segmentation divides an image into meaningful regions, often identifying the pixels associated with particular objects or areas. Unlike basic image classification, which generally assigns a category to an entire image, segmentation provides more detailed information about where specific regions or objects exist within the image. Object detection can identify objects and their locations, commonly using bounding boxes, while segmentation can provide more precise pixel-level boundaries. Regression predicts numerical values and is not primarily a computer vision region-identification technique. Therefore, when the requirement is to separate an image into meaningful regions or identify objects at the pixel level, segmentation is the appropriate computer vision task.

Question 46. Which Azure AI capability can summarize the main ideas from a long text?

1. Text summarization
2. Object detection
3. Speech recognition
4. Image classification

Correct Answer: 1. Text summarization

Explanation:

Text summarization is a natural language processing capability used to produce a shorter representation of a longer piece of text while retaining important information. It can help users quickly understand documents, reports, articles, support conversations, or other lengthy content without reading every sentence. Depending on the implementation, summarization may produce extractive results based on existing text or generate a more concise summary. Object detection and image classification process visual information, while speech recognition converts spoken audio into text. Therefore, when an application needs to identify and present the main ideas from a long written document, text summarization is the appropriate AI capability.

Question 47. What is the main purpose of a machine learning training dataset?

1. To train the model to learn patterns from data
2. To permanently store production predictions
3. To configure DNS records
4. To monitor network traffic

Correct Answer: 1. To train the model to learn patterns from data

Explanation:

A training dataset provides examples that a machine learning algorithm uses to learn patterns and relationships relevant to the prediction task. In supervised learning, the training examples generally contain input features together with known target labels or values. The learning algorithm adjusts model parameters based on these examples so that the resulting model can make predictions on new data. A training dataset is not primarily a production prediction store, DNS configuration source, or network-monitoring mechanism. After training, separate validation or test data can be used to evaluate model behavior. Therefore, the primary purpose of a training dataset is to provide examples from which the machine learning model can learn.

Question 48. Which Azure Machine Learning capability helps automate a sequence of data preparation and model training steps?

1. Pipeline
2. DNS zone
3. Storage account key
4. Network security group

Correct Answer: 1. Pipeline

Explanation:

An Azure Machine Learning pipeline can organize multiple machine learning steps into a repeatable workflow. A pipeline may include activities such as data preparation, feature processing, model training, evaluation, and other supported operations. Automating these steps can improve consistency and make machine learning workflows easier to reproduce and manage. Pipelines are particularly useful when a process needs to be executed repeatedly as data or models change. DNS zones manage name resolution, storage account keys provide access credentials, and network security groups control network traffic. Therefore, when the requirement is to automate a sequence of machine learning workflow steps, an Azure Machine Learning pipeline is the appropriate capability.

Question 49. Which machine learning problem predicts whether a customer will cancel a subscription?

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

Correct Answer: 2. Classification

Explanation:

Predicting whether a customer will cancel a subscription is commonly treated as a classification problem because the target can be represented as discrete categories, such as cancel or not cancel. A supervised classification model can learn from historical customer records containing relevant features and known outcomes. After training, the model can classify new customers according to the learned patterns. Regression is intended for continuous numerical predictions, clustering discovers groups without predefined labels, and dimensionality reduction transforms data into a smaller representation. Therefore, if the expected output is a discrete outcome such as whether a customer will cancel a subscription, classification is the appropriate machine learning problem type.

Question 50. Which machine learning problem predicts the expected monthly electricity consumption of a building?

1. Classification
2. Clustering
3. Regression
4. Object detection

Correct Answer: 3. Regression

Explanation:

Predicting expected monthly electricity consumption involves estimating a numerical quantity, such as kilowatt-hours consumed during a future period. This makes regression an appropriate supervised machine learning approach. The model can learn relationships between historical consumption and relevant features such as building characteristics, weather information, occupancy, and previous usage. It can then produce a numerical prediction for a new period. Classification is used for discrete categories, clustering groups similar observations without predefined labels, and object detection identifies objects within images. Therefore, when the output is a continuous numerical value representing expected electricity consumption, regression is the appropriate machine learning problem type.

Question 51. Which Azure Machine Learning component is used to store and organize reusable machine learning data assets?

1. Data asset
2. Virtual network
3. DNS zone
4. Load balancer

Correct Answer: 1. Data asset

Explanation:

Azure Machine Learning data assets provide a way to register and manage references to data used by machine learning workflows. Organizing data as managed assets can make it easier for teams to reuse datasets and maintain consistent references across training and other machine learning jobs. Data assets can represent data stored in supported Azure and other accessible locations without necessarily requiring the data itself to be copied into the machine learning workspace. Virtual networks provide network isolation, DNS zones manage name resolution, and load balancers distribute traffic. Therefore, a data asset is the appropriate Azure Machine Learning concept for organizing reusable data used by machine learning workflows.

Question 52. Which Azure Machine Learning resource provides an isolated environment for organizing machine learning assets and experiments?

1. Storage container
2. Machine learning workspace
3. DNS server
4. Virtual machine scale set

Correct Answer: 2. Machine learning workspace

Explanation:

An Azure Machine Learning workspace provides a centralized environment for organizing machine learning resources and assets. Depending on the workflow, a workspace can contain or reference items such as models, data assets, experiments, jobs, compute resources, and endpoints. It provides a management boundary that helps teams organize their machine learning activities and associated resources. A storage container is primarily used to organize stored objects, a DNS server handles name resolution, and a virtual machine scale set manages groups of virtual machines. Therefore, when an organization needs a centralized environment for managing machine learning assets and experiments, an Azure Machine Learning workspace is the appropriate resource.

Question 53. Which Azure Machine Learning resource provides computing capacity for running training jobs?

1. Compute
2. Key Vault
3. DNS zone
4. Application Gateway

Correct Answer: 1. Compute

Explanation:

Azure Machine Learning compute resources provide processing capacity for machine learning workloads such as model training and other jobs. Depending on the scenario, organizations can use managed compute resources or attach suitable existing resources to support their workflows. Selecting appropriate compute capacity is important because training requirements can vary significantly based on dataset size, model complexity, and workload duration. Key Vault is intended for secure secret and key management, DNS zones provide name resolution, and Application Gateway is a web traffic management service. Therefore, when a machine learning job requires processing resources for training or related workloads, Azure Machine Learning compute is the relevant resource.

Question 54. Which concept describes the settings chosen before training that control how a machine learning algorithm learns?

1. Labels
2. Features
3. Hyperparameters
4. Predictions

Correct Answer: 3. Hyperparameters

Explanation:

Hyperparameters are configuration values selected before or during the model-training process that influence how a machine learning algorithm learns. Examples can include learning rate, number of trees in certain ensemble models, maximum tree depth, batch size, and other algorithm-specific settings. Hyperparameters are different from model parameters, which are generally learned from the training data. Features are input variables, labels represent target values in supervised learning, and predictions are outputs produced after a model is applied to data. Therefore, settings that are selected to control aspects of the learning process are referred to as hyperparameters.

Question 55. Which machine learning technique can reduce the number of input dimensions while retaining important information?

1. Classification
2. Dimensionality reduction
3. Sentiment analysis
4. Object detection

Correct Answer: 2. Dimensionality reduction

Explanation:

Dimensionality reduction transforms data containing many variables into a representation with fewer dimensions while attempting to preserve important information or structure. Reducing the number of dimensions can help simplify datasets, reduce computational requirements, and sometimes make patterns easier to visualize or analyze. Techniques such as principal component analysis are commonly associated with dimensionality reduction. Classification predicts categories, sentiment analysis evaluates language sentiment, and object detection identifies objects within images. Dimensionality reduction does not necessarily mean simply deleting arbitrary features; appropriate techniques transform or select information according to a defined method. Therefore, dimensionality reduction is the machine learning technique described.

Question 56. Which responsible AI principle focuses on designing AI systems that can be used by people with different abilities and backgrounds?

1. Inclusiveness
2. Transparency
3. Reliability and safety
4. Accountability

Correct Answer: 1. Inclusiveness

Explanation:

Inclusiveness is a responsible AI principle focused on designing systems that can serve people with diverse abilities, backgrounds, and circumstances. An inclusive AI solution should consider accessibility and different user needs during design, development, testing, and deployment. For example, an application may need to account for users with different communication methods or accessibility requirements. Transparency concerns making AI behavior understandable, reliability and safety concern dependable and safe operation, and accountability concerns responsibility for AI outcomes. Therefore, when the requirement is to design an AI system that can be used effectively by people with diverse needs and abilities, inclusiveness is the relevant responsible AI principle.

Question 57. Which responsible AI principle emphasizes that organizations remain responsible for the decisions and outcomes of their AI systems?

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

Correct Answer: 2. Accountability

Explanation:

Accountability means that people and organizations remain responsible for the development, deployment, monitoring, and outcomes of AI systems. Organizations should establish appropriate processes for governance, oversight, risk management, and human responsibility rather than treating an AI system as independently responsible for its decisions. Accountability can include documenting decisions, monitoring system behavior, defining ownership, and establishing procedures for addressing problems. Fairness focuses on equitable treatment, inclusiveness considers diverse user needs, and privacy concerns protection of personal information. Therefore, the responsible AI principle that emphasizes organizational responsibility for AI decisions and outcomes is accountability.

Question 58. Which Azure AI service is most appropriate for extracting structured information from invoices and receipts?

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

Correct Answer: 1. Azure AI Document Intelligence

Explanation:

Azure AI Document Intelligence is designed to extract information from documents using document analysis and recognition capabilities. It can process supported documents such as invoices, receipts, forms, and other structured or semi-structured content and return information in a structured form. This can reduce the need to manually enter data from business documents and can support automated processing workflows. Azure AI Speech focuses on spoken language, Azure AI Search provides search and retrieval capabilities, and Azure DNS provides domain-name resolution. Therefore, when an application needs to extract structured information such as invoice fields or receipt details from documents, Azure AI Document Intelligence is the appropriate service.

Question 59. Which Azure AI service can answer questions using information retrieved from an organization’s indexed documents?

1. Azure AI Search
2. Azure AI Speech
3. Azure AI Vision
4. Azure Firewall

Correct Answer: 1. Azure AI Search

Explanation:

Azure AI Search can index organizational content and provide search and retrieval capabilities that help applications find relevant information from documents and other supported sources. In generative AI architectures, retrieved information can be supplied as grounding context to a language model so that generated responses can be based on organizational content. The search service itself focuses on finding and retrieving relevant information rather than independently replacing a generative model. Azure AI Speech processes spoken audio, Azure AI Vision analyzes images, and Azure Firewall provides network security. Therefore, Azure AI Search is an appropriate service for retrieving relevant information from indexed organizational documents for question-answering workflows.

Question 60. Which Azure AI capability helps identify and filter potentially harmful or unsafe content?

1. Azure AI Content Safety
2. Azure AI Speech
3. Azure AI Vision only
4. Azure DNS

Correct Answer: 1. Azure AI Content Safety

Explanation:

Azure AI Content Safety provides capabilities designed to help detect and manage potentially harmful content in AI applications. Depending on the supported feature, it can analyze content categories associated with unsafe or harmful material and help organizations apply appropriate filtering or moderation controls. This can be important when applications accept user-generated content or generate responses using AI models. Azure AI Speech focuses on speech processing, Azure AI Vision provides computer vision capabilities, and Azure DNS handles domain-name resolution. Therefore, when an AI application needs capabilities for identifying and managing potentially harmful content, Azure AI Content Safety is the appropriate Azure AI service.