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

 

Question 21. Which Azure service provides natural language processing capabilities for analyzing text?

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

Correct Answer: 2. Azure AI Language

Explanation:

Azure AI Language provides a range of natural language processing capabilities for analyzing and understanding text. Depending on the feature used, it can support tasks such as sentiment analysis, key phrase extraction, entity recognition, language detection, summarization, and question answering. These capabilities allow applications to derive meaningful information from written language without requiring developers to build every language-processing model from scratch. Azure AI Vision focuses on image analysis, Azure AI Speech handles speech-related scenarios, and Azure Machine Learning provides a broader platform for machine learning development. Therefore, Azure AI Language is the appropriate service for text-based natural language processing.

Question 22. Which Azure AI service is designed to convert spoken audio into text?

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

Correct Answer: 1. Azure AI Speech

Explanation:

Azure AI Speech provides speech-processing capabilities, including speech-to-text conversion. Speech recognition allows an application to receive spoken audio and produce a text representation of what was said. This capability can be useful in applications such as transcription, voice-controlled interfaces, meeting analysis, and accessibility solutions. Azure AI Speech also supports other speech scenarios, including text-to-speech and speech-related features depending on the service capabilities being used. Azure AI Vision is designed for visual content, Azure AI Language focuses on text understanding, and Azure AI Search supports search and retrieval. Therefore, Azure AI Speech is the appropriate service for converting spoken audio into text.

Question 23. Which AI workload involves identifying whether an image contains a specific object or category?

1. Speech synthesis
2. Text summarization
3. Image classification
4. Regression

Correct Answer: 3. Image classification

Explanation:

Image classification is a computer vision workload in which an AI model assigns one or more predefined categories to an image based on its visual content. For example, an application might classify images as containing cats, dogs, cars, or other known categories. The model learns from training examples and then uses learned visual patterns to classify new images. Object detection is different because it can identify objects and their locations within an image, often using bounding boxes. Speech synthesis converts text into spoken audio, text summarization condenses written content, and regression predicts numerical values. Therefore, identifying an image category is an image classification task.

Question 24. Which computer vision task identifies objects and their locations within an image?

1. Object detection
2. Sentiment analysis
3. Speech recognition
4. Text classification

Correct Answer: 1. Object detection

Explanation:

Object detection is a computer vision task that identifies objects within an image and determines where those objects are located. A typical object-detection result can include the detected object’s category together with a bounding box showing its position and dimensions. This makes object detection useful for scenarios such as identifying vehicles in traffic images, detecting products on shelves, or locating people and equipment in visual environments. Image classification generally assigns categories to an entire image without necessarily locating individual objects. Sentiment analysis and text classification work with language, while speech recognition processes spoken audio. Therefore, object detection is the correct computer vision task.

Question 25. Which Azure AI capability can determine whether a text review expresses positive or negative sentiment?

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

Correct Answer: 2. Sentiment analysis

Explanation:

Sentiment analysis is a natural language processing capability that evaluates text to determine the expressed sentiment, such as positive, negative, or neutral. It can be applied to customer reviews, survey responses, social media content, support tickets, and other textual information. For example, an organization could analyze thousands of customer comments to identify general satisfaction trends. Sentiment analysis does not necessarily determine the factual correctness of a statement; it focuses on the emotional or opinion-related polarity expressed in the text. Object detection and image classification process visual information, while speech synthesis converts text into spoken audio. Therefore, sentiment analysis is the appropriate capability for evaluating the sentiment of a text review.

Question 26. Which Azure AI Language capability identifies important terms or concepts in a document?

1. Key phrase extraction
2. Speech recognition
3. Object detection
4. Image classification

Correct Answer: 1. Key phrase extraction

Explanation:

Key phrase extraction identifies important words or phrases that represent the main concepts in a body of text. It can help applications quickly summarize the major topics discussed in documents, reviews, support tickets, or other textual content. For example, a customer complaint containing several sentences might produce key phrases related to delivery delays, damaged products, and customer service. This capability is useful when organizations need to process large amounts of text and identify major themes automatically. Speech recognition converts audio into text, while object detection and image classification analyze visual content. Therefore, key phrase extraction is the appropriate Azure AI Language capability.

Question 27. Which Azure AI Language capability can identify people, organizations, locations, and other named entities in text?

1. Entity recognition
2. Image segmentation
3. Speech synthesis
4. Regression

Correct Answer: 1. Entity recognition

Explanation:

Entity recognition identifies named entities and other meaningful references within text. Depending on the supported model and configuration, entities can include people, organizations, locations, dates, quantities, and other recognized categories. This capability helps applications transform unstructured text into structured information that can be searched, analyzed, or used in downstream workflows. For example, a document-processing application could extract organization names and locations from business reports. Image segmentation is a computer vision task, speech synthesis converts text into audio, and regression predicts numerical values. Therefore, entity recognition is the appropriate natural language processing capability for identifying named entities within text.

Question 28. Which AI workload converts written text into natural-sounding spoken audio?

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

Correct Answer: 3. Text-to-speech

Explanation:

Text-to-speech converts written text into spoken audio using synthesized speech. This capability can be used to create voice assistants, accessibility features, automated announcements, narration systems, and other applications where written information needs to be presented as audio. Modern speech services can provide different voices, languages, and speaking characteristics depending on the available configuration. Speech-to-text performs the opposite general transformation by converting spoken audio into text. Text classification categorizes written content, while object detection analyzes visual content. Therefore, when an application needs to turn written text into spoken audio, text-to-speech is the appropriate AI workload.

Question 29. Which responsible AI principle focuses on ensuring an AI system performs consistently and safely under expected conditions?

1. Reliability and safety
2. Fairness
3. Transparency
4. Privacy

Correct Answer: 1. Reliability and safety

Explanation:

Reliability and safety focus on ensuring that an AI system operates consistently and safely within the conditions for which it was designed. AI solutions should be tested and monitored to identify unexpected behavior, failures, or harmful outcomes. Appropriate safeguards can help reduce risks when systems encounter unusual inputs or conditions. Fairness addresses equitable treatment across people or groups, transparency focuses on understanding how AI systems work and produce results, and privacy concerns the appropriate handling and protection of personal information. Therefore, when the requirement concerns dependable and safe AI operation under expected conditions, reliability and safety is the relevant responsible AI principle.

Question 30. Which responsible AI principle focuses on protecting personal information used by an AI system?

1. Inclusiveness
2. Privacy and security
3. Transparency
4. Reliability

Correct Answer: 2. Privacy and security

Explanation:

Privacy and security focus on protecting personal information and other sensitive data used or processed by AI systems. Organizations should consider how data is collected, stored, accessed, processed, and retained and should apply appropriate security controls. AI applications may process personal information such as names, contact details, financial information, or other sensitive records, making responsible data handling important. Inclusiveness concerns designing systems that work for people with diverse needs, transparency concerns understandable AI behavior, and reliability concerns dependable operation. Therefore, the responsible AI principle most directly associated with protecting personal information and securing AI data is privacy and security.

Question 31. Which type of generative AI model is commonly used to generate and understand natural language?

1. Large language model
2. Object detection model
3. Regression model
4. Clustering algorithm

Correct Answer: 1. Large language model

Explanation:

A large language model, or LLM, is a type of AI model designed to process and generate natural language. Depending on its capabilities, an LLM can support tasks such as answering questions, summarizing information, generating text, translating content, extracting information, and assisting with code. These models are trained on large collections of data and learn statistical patterns in language. Object detection models focus on visual objects, regression models predict numerical values, and clustering algorithms group similar data points. Therefore, a large language model is the type of generative AI model commonly associated with understanding and generating natural language.

Question 32. What is a prompt in a generative AI application?

1. A database backup
2. An instruction or input provided to the AI model
3. A network firewall rule
4. A machine learning hardware component

Correct Answer: 2. An instruction or input provided to the AI model

Explanation:

A prompt is the input or instruction provided to a generative AI model to guide the response or content it produces. A prompt can be a question, command, description, context, or combination of instructions and information. The quality and specificity of a prompt can influence the relevance and usefulness of the model’s output. For example, a developer might provide instructions that specify the desired format, audience, tone, and task. A prompt is not a database backup, firewall rule, or hardware component. Therefore, an instruction or input supplied to a generative AI model is called a prompt.

Question 33. Which prompt technique provides examples to guide a generative AI model’s response?

1. Zero-shot prompting
2. Few-shot prompting
3. Data encryption
4. Image segmentation

Correct Answer: 2. Few-shot prompting

Explanation:

Few-shot prompting provides a generative AI model with a small number of examples demonstrating the desired type of input and output. These examples can help the model understand the expected pattern, format, or classification behavior without requiring the underlying model to be retrained. For example, a prompt might provide several customer comments paired with sentiment labels and then ask the model to classify another comment. Zero-shot prompting provides instructions without examples. Data encryption protects information, while image segmentation is a computer vision technique. Therefore, providing examples in a prompt to guide the model is known as few-shot prompting.

Question 34. Which Azure service is commonly used to provide enterprise search over indexed content?

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

Correct Answer: 1. Azure AI Search

Explanation:

Azure AI Search provides search capabilities that allow applications to index and retrieve information from supported content sources. It can be used to create search experiences over organizational documents, websites, application data, and other content. Search indexes can support capabilities such as keyword search, filtering, sorting, and other supported retrieval features. Azure AI Search can also participate in generative AI solutions where retrieved organizational information is supplied as grounding context. Azure AI Speech focuses on audio, Azure AI Vision processes images, and Azure DNS provides name resolution. Therefore, Azure AI Search is the appropriate service for enterprise search over indexed content.

Question 35. What is the purpose of grounding in a generative AI solution?

1. To provide relevant external information to improve generated responses
2. To increase network bandwidth
3. To replace all model training
4. To encrypt every prompt

Correct Answer: 1. To provide relevant external information to improve generated responses

Explanation:

Grounding provides a generative AI model with relevant information from trusted external sources so that its responses can be better connected to the information available to an organization or application. For example, a conversational application might retrieve current information from a company knowledge base and provide that information as context before asking a language model to formulate an answer. Grounding can help reduce reliance on information contained only in the model’s training data. It does not increase network bandwidth, automatically replace model training, or simply encrypt prompts. Therefore, providing relevant external information as context is a primary purpose of grounding.

Question 36. Which Azure service can be used to store unstructured documents that may later be indexed for AI search?

1. Azure Blob Storage
2. Azure DNS
3. Azure Firewall
4. Azure Load Balancer

Correct Answer: 1. Azure Blob Storage

Explanation:

Azure Blob Storage is designed to store large amounts of unstructured data, including documents, images, text files, and other content. Such content can be used as a source for downstream processing, including search and AI workflows, depending on the architecture. An organization might store a collection of PDF documents or other business files in blob containers and then process or index them for retrieval. Azure DNS provides name resolution, Azure Firewall provides network security, and Azure Load Balancer distributes network traffic. Therefore, Azure Blob Storage is a suitable service for storing unstructured documents that may later participate in AI-powered search workflows.

Question 37. Which machine learning concept describes a model’s ability to perform well on new, unseen data?

1. Generalization
2. Overfitting
3. Memorization
4. Data duplication

Correct Answer: 1. Generalization

Explanation:

Generalization describes a machine learning model’s ability to perform effectively on new data that was not included in its training examples. A model that generalizes well has learned useful patterns rather than simply memorizing the training dataset. Good generalization is an important goal of machine learning because production applications typically encounter data that differs from the exact examples used during training. Overfitting occurs when a model learns training-specific patterns too closely and performs poorly on unseen data. Memorization and data duplication do not describe the desired predictive behavior. Therefore, the ability to perform well on previously unseen data is called generalization.

Question 38. Which metric is commonly used to evaluate a regression model’s prediction error?

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

Correct Answer: 2. Mean absolute error

Explanation:

Mean absolute error, or MAE, is a commonly used metric for evaluating regression models. It calculates the average absolute difference between the model’s predicted values and the actual target values. Because the errors are converted to absolute values before averaging, positive and negative errors do not cancel each other out. MAE is expressed in the same general units as the target variable, which can make it relatively easy to interpret. Accuracy is commonly associated with classification, while precision and recall are also classification-oriented metrics. Therefore, mean absolute error is an appropriate metric for measuring prediction error in a regression model.

Question 39. Which metric measures the proportion of predicted positive results that are actually positive?

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

Correct Answer: 3. Precision

Explanation:

Precision measures the proportion of predicted positive results that are actually positive. It is calculated as true positives divided by the sum of true positives and false positives. Precision is especially useful when false-positive predictions have a significant cost. For example, in a system that flags transactions as potentially fraudulent, high precision means that a larger proportion of flagged transactions are actually positive cases. Recall instead measures how many of the actual positive cases were identified. Accuracy measures overall correct predictions, while mean absolute error is primarily used for regression. Therefore, precision is the metric described in the question.

Question 40. Which metric measures how many actual positive cases a classification model successfully identifies?

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

Correct Answer: 1. Recall

Explanation:

Recall measures the proportion of actual positive cases that a classification model successfully identifies. It is calculated as true positives divided by the total number of actual positive cases, which includes both true positives and false negatives. Recall is particularly important when missing a positive case can have significant consequences. For example, in a screening system, high recall can help ensure that more of the relevant cases are identified for further review. Precision instead focuses on how many predicted positives are actually positive. Mean squared error is commonly used for regression, while accuracy measures overall correct predictions. Therefore, recall is the correct metric.