Microsoft Azure AI Fundamentals Practice Test (AI-900)

The Codex family of models in Azure OpenAI Service was specifically trained on a massive corpus of public code and natural language to translate descriptions into executable code. While newer models like GPT-4 also have strong coding capabilities, the term 'Codex' refers to the models originally specialized for this purpose, as reflected in the AI-900 exam objectives.

The Azure AI Vision service provides extensive image analysis capabilities, including identifying objects within images and extracting text through optical character recognition (OCR).

Azure AI Face service is specialized for detecting and analyzing human faces, but doesn't support general object identification or text extraction.

Azure AI Document Intelligence is designed to extract structured data from forms and documents but is less suitable for general image analysis tasks.

Azure Speech to Text service is used for converting spoken language into text and does not process images.

The Azure AI Face Detection service is designed specifically for detecting human faces and analyzing facial features like age, emotion and gender. It provides advanced facial recognition capabilities tailored for these tasks.

The Azure AI Vision service offers general image analysis, it does not specialize in detailed facial attribute analysis.

Azure AI Form Recognizer is used for extracting text from forms and documents.

Azure AI Speech service processes and recognizes spoken language, neither of which address facial detection or analysis.

Using a training dataset that covers all relevant demographic groups exposes the model to varied examples and reduces the chance that it learns patterns that favor or exclude particular populations.

Reducing the sampling temperature only makes outputs more deterministic; it does not address underlying bias in the learned parameters.

Removing outliers that correspond to minority groups shrinks representation and tends to worsen bias.

Simply increasing the depth or parameter count of the network can even amplify existing bias if the data remain unbalanced.

Content Synthesis tasks - This is the correct answer. Content synthesis tasks involve generating new content that is similar to the examples the model has been trained on. This is typically seen in generative models, where the system can create new, original content such as text, images, or music based on the patterns it has learned.

Anomaly Detection focuses on identifying outliers or unusual patterns in data, not on generating new content. It is used for detecting abnormal behavior or data points.

Data Classification involves categorizing data into predefined classes or categories, such as labeling emails as spam or not. It does not involve generating new content.

Regression Analysis is used to predict numerical values based on historical data, such as forecasting sales or stock prices. It is not focused on generating new content.

Clustering - This is the correct answer. Clustering is an unsupervised machine learning technique used to group data into segments based on similarities without requiring labeled data. It is ideal for grouping customers into segments based on their behavior.

Regression is used for predicting continuous numerical values based on input features, not for grouping data or segmenting customers.

Classification is a supervised learning technique where the goal is to categorize data into predefined classes. It requires labeled data, which is not available in this case.

Time Series Analysis is used to analyze data that is collected over time (e.g., stock prices, sales data) to identify trends or patterns. It is not focused on segmenting data into groups based on behavior.

DALL·E is an image generation model available in Azure OpenAI Service that creates images from text descriptions, allowing users to generate visual content based on textual inputs.

GPT-3 and ChatGPT are language models designed for generating and understanding human-like text, not images.

Codex is tailored for code generation and completion tasks and does not generate images.

The design decisions focus on Inclusiveness. Microsoft's Inclusiveness principle states that AI systems should "empower everyone and engage all people, regardless of their backgrounds" and be inclusive "for people of all abilities." Adding screen-reader friendly labels, keyboard navigation, and descriptive alt-text removes barriers for users with disabilities, directly addressing this principle. The other options target different Responsible AI areas: Fairness concerns bias and equal treatment, Transparency is about making the system understandable, and Accountability relates to human oversight.

Deep learning techniques utilize neural networks with multiple layers (deep neural networks) to automatically extract high-level features from raw data. This ability to perform automatic feature extraction eliminates the need for manual feature engineering, which is often required in traditional machine learning methods. The other options are incorrect because deep learning models typically require large datasets to perform effectively, they can process both structured and unstructured data, and they are generally less interpretable due to their complex architectures.

Azure Face can locate faces, return landmark coordinates, and perform biometric verification or identification. These are all documented features of the service . It does not create a detailed 3-D mesh model suitable for rendering in a game engine; developers who need a 3-D facial mesh must rely on specialized 3-D reconstruction libraries or hardware such as depth cameras.

Knowledge mining is the appropriate AI workload because it orchestrates services such as OCR, natural language processing, and search indexing to ingest, enrich, and make large, heterogeneous document collections easily searchable.

Natural language processing focuses on understanding and generating language from already-available text but does not, by itself, provide the pipeline to ingest multiple file formats or build a search index.

Computer Vision can extract text from images and scanned pages through OCR, but on its own it lacks the enrichment and indexing steps required to turn a mixed-format content repository into a searchable knowledge base.

Content personalization tailors experiences to individual users' preferences and does not address the need to search across a document repository.

Analyzing images and extracting visual information - This is the correct answer. The Azure AI Vision service provides developers with capabilities to analyze images and extract visual information. This includes tasks such as image classification, object detection, and optical character recognition (OCR), helping developers to gain insights from images.

Transcribing spoken language into text - This is the functionality of speech recognition services, not the Azure AI Vision service. Speech recognition is focused on converting spoken language into written text.

Translating text between different languages - This is the functionality of Azure Translator service, which is designed for translating text between languages, not related to image analysis.

Performing sentiment analysis on text data - This is the functionality of text analytics services, which are used for tasks like sentiment analysis, key phrase extraction, and language detection, but it is not related to image analysis.

The Azure AI Language service provides pre-built natural language processing (NLP) capabilities such as language detection, key phrase extraction, named entity recognition, and sentiment analysis. It allows developers to integrate these features into their applications without having to develop custom machine learning models.

Other services like Azure AI Speech service focus on speech-related functionalities, Azure Cognitive Search provides indexing and search capabilities, and Azure Machine Learning Service is for building and training custom machine learning models.

They can generate new data instances similar to the training data - This is the correct answer. Generative AI models are designed to create new data instances that resemble the training data. This makes them ideal for generating synthetic data to augment testing datasets, ensuring variety and realism in the data used for testing.

They can classify data into specific categories with high precision - This feature is characteristic of classification models, which categorize data into predefined labels but do not generate new data instances.

They can reduce data dimensionality while retaining key features - This describes dimensionality reduction techniques (such as PCA), which focus on simplifying the data without losing important features. It is not related to generating new data instances.

They can identify anomalies by learning normal data patterns - This is a feature of anomaly detection models, which are used to identify outliers or unusual patterns in data, not for generating synthetic data.

Azure OpenAI Service's natural language generation capability allows users to generate human-like text content such as articles, summaries, and conversational responses based on input prompts.

Other options like data visualization, speech-to-text transcription, and code compilation do not facilitate the creation of such text-based content.

Classification is used when the goal is to assign data points to discrete categories based on input features. In this case, the hospital aims to categorize patients as having the disease or not, which is a typical classification problem.

Regression is meant for predicting continuous numerical values, such as prices or temperatures.

Clustering is where data points are grouped based on similarity without predefined labels, which doesn't fit the need to assign patients to known categories.

Reinforcement Learning involves learning optimal actions through trial and error, which is not applicable here.

Azure OpenAI Service supports code generation, which enables developers to create software components based on natural language descriptions. Using models like Codex, developers can input text prompts describing the desired functionality, and the service generates the corresponding code. This greatly simplifies coding tasks and accelerates development.

Image synthesis refers to generating images from text descriptions, not software components. Azure OpenAI Service offers code generation but does not synthesize images for creating software components; image synthesis is a separate capability found in tools like DALL-E.

Data analysis involves examining and interpreting data to derive insights, which is unrelated to creating software components from text descriptions. While Azure OpenAI can assist in querying data, the service’s code generation capability specifically translates natural language into code rather than performing data analysis.

Speech-to-text conversion transcribes spoken language into text, which does not assist in generating software components. Azure offers separate speech services for voice transcription, while code generation is a distinct feature that translates text-based prompts into code outputs.

Discovering customer segments with similar purchasing behaviors for marketing purposes - This is the correct answer. The scenario where a machine learning technique finds patterns in data without relying on labeled outcomes is best addressed by clustering. In this case, the goal is to segment customers based on their purchasing behaviors, which is an unsupervised learning task that does not require labeled data.

Predicting housing prices based on features like size and location - This task typically requires supervised learning, where the model learns from labeled data, for example, by using historical housing prices to make predictions on new data.

Determining if a transaction is fraudulent based on past labeled data - Fraud detection is a supervised learning task where labeled data (fraudulent vs. legitimate transactions) is used to train a model to classify future transactions.

Forecasting future stock prices based on historical data - While stock price forecasting may involve time series analysis, it is also typically a supervised learning task that uses historical data with known outcomes to predict future trends.

Regression - This is the correct answer. Regression is the appropriate machine learning technique for predicting continuous numerical values, such as the delivery duration for packages. The company can use regression to model the relationship between the input factors (distance, traffic conditions, weather) and the target variable (delivery duration).

Classification is used to predict categorical outcomes, not continuous values like delivery duration. Since the goal is to predict a numeric value, regression is the more suitable approach.

Clustering is an unsupervised learning technique that groups data based on similarities, but it is not used for predicting specific outcomes like delivery duration. It's useful for finding patterns, but not for making specific numerical predictions.

Reinforcement Learning is typically used for decision-making tasks where an agent learns by interacting with an environment for example optimizing actions over time. It is not the best choice for predicting continuous values like delivery duration, which is more suited to regression.

Azure AI Document Intelligence is specifically designed to extract structured data from scanned documents like invoices and receipts. It uses machine learning models to identify and extract key-value pairs, text, and tables, transforming unstructured documents into structured data.

Azure AI Search is used for indexing and searching over large sets of data but is not the primary service for extracting structured data from document layouts.

Azure AI Language processes unstructured text to detect sentiment, key phrases, and entities but doesn't work directly with the layout and structure of scanned documents.

Azure Computer Vision OCR extracts text from images but doesn't inherently structure the data or extract key-value pairs as needed for invoices and receipts.