Microsoft Azure AI Fundamentals General 2 — Questions and Answers
Question 1: Which Azure service provides a unified interface for deploying and managing large language models from multiple providers?
- Azure Machine Learning
- Azure OpenAI Service
- Azure AI Foundry (Correct answer)
- Azure Cognitive Search
Correct answer: Azure AI Foundry
Azure AI Foundry (formerly Azure AI Studio) is the unified platform for discovering, deploying, and managing AI models from Microsoft and third-party providers.
Question 2: In the context of responsible AI, what does 'fairness' primarily require?
- All users receive identical outputs
- AI systems do not produce biased outcomes that disadvantage specific groups (Correct answer)
- Models are trained on balanced dataset sizes
- Predictions are equally fast for all demographics
Correct answer: AI systems do not produce biased outcomes that disadvantage specific groups
Fairness in responsible AI means that AI systems should not produce discriminatory outcomes that unfairly disadvantage people based on characteristics such as race, gender, or age.
Question 3: A company wants to detect whether a product image contains a defect on the assembly line. Which Azure AI capability is most appropriate?
- Azure AI Language
- Azure AI Vision custom model (Correct answer)
- Azure AI Speech
- Azure AI Translator
Correct answer: Azure AI Vision custom model
Azure AI Vision supports custom image classification and object detection models that can be trained to identify specific defects in product images.
Question 4: What is the purpose of 'grounding' in generative AI applications?
- Reducing the model's response length
- Connecting model outputs to verified, real-world data sources (Correct answer)
- Training the model on domain-specific data
- Limiting the number of API calls per session
Correct answer: Connecting model outputs to verified, real-world data sources
Grounding connects a generative AI model's responses to authoritative, real-world data sources to reduce hallucinations and improve factual accuracy.
Question 5: Which metric is most commonly used to evaluate the performance of a classification model?
- Mean Absolute Error
- R-squared
- Accuracy (Correct answer)
- Root Mean Squared Error
Correct answer: Accuracy
Accuracy, which measures the proportion of correctly classified instances, is the most commonly used evaluation metric for classification models.
Question 6: Azure AI Content Safety is designed to detect which type of content?
- Grammatical errors in text
- Harmful or inappropriate content such as hate speech and violence (Correct answer)
- Personally identifiable information (PII)
- Duplicate content across documents
Correct answer: Harmful or inappropriate content such as hate speech and violence
Azure AI Content Safety analyzes text and images to detect harmful content categories including hate speech, violence, sexual content, and self-harm.
Question 7: Which statement best describes the difference between AI and machine learning?
- AI and machine learning are identical concepts
- Machine learning is a subset of AI focused on learning from data (Correct answer)
- AI is a subset of machine learning
- Machine learning requires no data to function
Correct answer: Machine learning is a subset of AI focused on learning from data
Machine learning is a subset of artificial intelligence where systems learn from data to improve their performance, while AI is the broader concept of machines simulating intelligence.
Microsoft Azure AI Fundamentals General 2
Which Azure service provides a unified interface for deploying and managing large language models from multiple providers?