Section 1
AI Foundations
Preview
Preview: AI workload selection and fundamentals
- 4 of 10 lesson topics
- 1 overview segment
- 3 core concepts
- 2 exam tips
Lesson Topics
- Predictive AI versus generative AI
- Prebuilt Azure AI services
- Custom Azure Machine Learning models
- Classification, regression, and anomaly detection
Overview
Start by identifying the input: tabular data, text, image, audio, document, or streaming telemetry. Then identify the output: category, number, extracted fields, generated content, search result, or anomaly alert.
Core Concepts
- **Predictive AI:** Uses data patterns to predict a category, value, anomaly, or next action. Choose it when the output is a prediction rather than generated text.
- **Generative AI:** Produces new content such as text, summaries, code, or images. Do not choose it when the scenario only needs a numeric forecast or class label.
- **Prebuilt Azure AI service:** A ready service for common workloads such as Vision, Language, Speech, Document Intelligence, Search, or Content Safety. Choose it when Microsoft already provides the needed capability.
Exam Tips
- Read the required output before picking a service.
- If the answer must be a number, think regression before generative AI.