Mixed Deck — All Microsoft Azure AI Fundamentals Topics Flashcards

Which Azure OpenAI model capability allows it to understand and respond to both text and images in the same prompt?

Answer: Multi-modal (vision) capability

Multi-modal models like GPT-4o can accept both text and image inputs, enabling use cases like image description, visual Q&A, and document analysis.

Which ML concept describes using a model trained on one task as a starting point for a different but related task?

Answer: Transfer learning

Transfer learning reuses knowledge from a pretrained model (e.g., ImageNet-trained CNN) and fine-tunes it for a new related task with less data.

In the context of Azure ML, what is a 'model registry'?

Answer: A centralized repository for storing, versioning, and managing trained ML models

The Azure ML model registry is a centralized store that tracks all trained models with versioning, metadata, and lineage information for governance and deployment.

Which capability allows Azure AI Language to identify the main topics discussed in a document?

Answer: Key phrase extraction

Key phrase extraction identifies the most important phrases that capture the main topics within a document.

What is an 'enrichment pipeline' in Azure AI Search?

Answer: A series of cognitive skill operations applied during document indexing to extract and enhance content

An enrichment pipeline in Azure AI Search applies cognitive skills sequentially to raw content during indexing, producing enriched, AI-annotated documents stored in the search index.

An organization needs to detect fraudulent financial transactions in real time. Which type of machine learning task is this?

Answer: Binary classification

Fraud detection predicts one of two outcomes (fraud or legitimate), making it a binary classification problem.

In Azure ML Designer, what is a 'pipeline' composed of?

Answer: Modules connected in a graph that represent data processing and model training steps

Azure ML Designer pipelines are visual graphs made up of connected modules (components) representing steps like data prep, training, and evaluation.

What is transfer learning in the context of AI model development?

Answer: Reusing a pre-trained model as a starting point and fine-tuning it for a new related task

Transfer learning leverages knowledge from a pre-trained model to accelerate training and improve performance on a new but related task.

Which kind of machine learning should you employ to forecast the quantity of gift cards that will be sold over the course of the upcoming month?

Answer: regression

Forecasting the quantity of gift cards sold over the upcoming month involves predicting a continuous numerical value. Regression models are specifically designed for tasks where the output is a numerical quantity. Classification is used for predicting discrete categories, and clustering is for grouping data, neither of which fits the requirement of predicting a specific numerical quantity.

Which of the following resources cannot be produced using Azure Machine Learning Studio?

Answer: Compute Balancers

Azure Machine Learning Studio provides various compute resources for training and deploying models, including Compute Instances (managed workstations), Compute Clusters (scalable training environments), and Inference Clusters (for deploying models). 'Compute Balancers' is not a standard resource type offered within Azure Machine Learning Studio.

Which responsible AI principle is most directly concerned with protecting personal information used by AI systems?

Answer: Privacy and security

The privacy and security principle requires that AI systems protect personal data and give users control over their information.

What is text-to-speech (TTS) synthesis in Azure AI Speech?

Answer: Converting written text into spoken audio output

Text-to-speech synthesis converts written text into natural-sounding spoken audio using neural voice models.

You're employed by a car dealership. Your employer requests that you give him forecast data. Will the new auto model succeed or fail? The new model sports a sunroof, improved seats, and many engine upgrades. You created a list of information about previous successful models, including details about their features and sales figures. What should you do to ensure the success of the new model during the pre-processing of the data stage?

Answer: Feature selection

To predict the success of a new car model, it's crucial to identify which characteristics (features like sunroof, seats, engine upgrades) from historical data are most influential. Feature selection is the process of choosing the most relevant features that contribute significantly to the prediction outcome. By focusing on these key features during data preprocessing, the model's accuracy and efficiency in forecasting success can be greatly improved.

In Azure Custom Vision, what is the minimum recommended number of images per tag for a good starting model?

Answer: 15 images

Azure Custom Vision recommends at least 15 images per tag as a starting point, though more images generally lead to better model accuracy.

Which of the following describes how artificial intelligence is advantageous?

Answer: All of the above

Artificial intelligence offers a wide range of benefits across various sectors. It significantly enhances security through advanced monitoring and threat detection, automates repetitive and complex tasks to improve efficiency and reduce human error, and drastically cuts down the time required to process information and solve problems. Therefore, all the listed points—providing security, making work easier, and reducing problem-solving time—are valid advantages of AI.

What does IPsec signify in the Azure platform?

Answer: Internet Protocol Security protocol suite

IPsec stands for Internet Protocol Security protocol suite. In the Azure platform, IPsec is a critical component for establishing secure, encrypted communication channels, particularly for Virtual Private Network (VPN) connections. It ensures data confidentiality, integrity, and authenticity when data travels between Azure virtual networks and on-premises networks, or between different Azure virtual networks.

How can you automatically extract text, key/value pairs, and table data from scanned documents? Which service should you use?

Answer: Form Recognizer

Azure Form Recognizer is an AI service that uses machine learning to identify and extract text, key/value pairs, and table data from documents. It is specifically designed to automate data extraction from various document types, including scanned forms, invoices, and receipts, making it ideal for this task.

Which Azure ML feature automatically selects the best algorithm and hyperparameters for a given dataset?

Answer: Automated ML (AutoML)

Automated ML (AutoML) in Azure Machine Learning iterates over multiple algorithms and hyperparameter settings to automatically find the best model for your data.

A business hires a group of customer care representatives to assist clients by phone and email. The business creates a webchat bot to offer pre-written responses to frequent client questions. What commercial advantages may the corporation anticipate from developing the webchat bot solution?

Answer: a reduced workload for the customer service agents

A webchat bot designed to answer frequent client questions can automate responses to common inquiries. This automation significantly reduces the number of routine tasks that human customer service agents need to handle. Consequently, it leads to a reduced workload for the agents, allowing them to focus on more complex or unique customer issues.

Which Azure ML compute target is best suited for running large-scale, distributed training jobs?

Answer: Azure ML Compute Cluster

Azure ML Compute Clusters scale out automatically to multiple nodes, making them ideal for distributed, large-scale training workloads.