AI-900 Exam Questions and Answers Lecture Notes

Introduction

• Overview of AI-900: Microsoft Azure AI Fundamentals
• Request to subscribe to YouTube channel for updates
• PDF available for download on shapingpixel.com

Exam Questions and Key Answers

Question 1: Ensuring Model Transparency

• Scenario: Build a machine learning model using Automated ML UI
• Objective: Meet Microsoft transparency principle
• Solution: Enable "Explain best model"
• Explanation: Transparency aids in understanding model predictions, especially in regulated industries like healthcare and banking.

Question 2: Data Transformation Modules in Azure ML

• Scenario: Identify data transformation modules
• Correct Options: Split data, Clean missing data
• Explanation: Essential for data preparation, includes tasks like cleaning and normalizing data.

Question 3: Benefits of Web Chat Bot

• Scenario: Company creating a web chat bot
• Main Benefit: Reduced workload for customer service agents

Question 4: Real-time Data Analysis for Factory Automation

• Scenario: Application needs to quickly identify sensor data issues
• Service: Azure Anomaly Detector
• Explanation: Detects anomalies in time-series data.

Question 5: Splitting Data for ML Model

• Scenario: Data splitting for training/evaluation
• Solution: Randomly split data into rows for training and evaluation

Question 6: Predicting Fuel Efficiency

• Scenario: Model to predict fuel efficiency
• Algorithm: Regression
• Explanation: Suitable for predicting continuous values.

Question 7: Confusion Matrix Analysis

• Scenario: Predict events using classification
• Specifics: Correct positives (11), False negatives (133)

Question 8: Building a FAQ-based Chat Bot

• Scenario: Use existing FAQ documentation
• Solution: Q&A Maker + Azure Bot Service

Question 9: Examples of Anomaly Detection

• True: Identifying suspicious sign-ins (Anomaly detection)
• False: Forecasting housing prices, predicting diabetes development

Question 10: Handling Unusual Values

• Principle: Reliability and safety
• Explanation: Ensures AI systems operate as designed.

Question 11: Language Understanding in AI

• Term for User Inquiry: Utterance

Question 12: AI Workloads and Scenarios

• Workloads: Conversational AI, Computer Vision, NLP
• Applications: Chatbots, photo content analysis, sentiment detection

Question 13: Inclusiveness Principle

• AI Design Objective: Empower everyone, including people with disabilities.

Question 14: Identifying Plant Species

• Service: Custom Vision for domain-specific image classification

Question 15: AI Principles and Descriptions

• Reliability & Safety: Systems should resist manipulation
• Accountability: Overriding AI decisions
• Privacy & Security: Control over data use

Additional Questions

• Discuss multiple scenarios, algorithms, and Azure services related to machine learning, data processing, chatbots, and AI principles.
• Focus on responsible AI practices and the application of Microsoft’s guiding principles such as transparency, inclusiveness, reliability, and safety.

Conclusion

• Understanding Azure AI services and responsible AI principles is crucial for AI-900 exam preparation.
• Emphasize ethical AI practices and technical proficiency in Azure machine learning tools.