A training note compares a hand-coded expert system, a decision tree trained on data, and an image classifier that learns with a multilayer neural network.
Explain how AI, machine learning and deep learning relate across the three examples.
Which statement is most appropriate?
AI is the broad concept, machine learning is one way to implement it, and deep learning is a type of machine learning.
Detailed explanation
AI is broad, while machine learning learns patterns from data as one implementation approach.
Machine learning learns patterns from training data or other examples.
Machine learning is commonly treated as part of AI, so the terms are not always identical.
Deep learning uses multilayer neural networks and is a type of machine learning.
Check official documentation
Compare the following points with the official documentation.
Read AWS's descriptions of machine learning and deep learning. For each of the three examples, identify which of the three categories apply.
Official references
What to confirm
The hand-coded expert system illustrates AI. The trained decision tree is AI and machine learning. The multilayer image classifier falls into all three categories.
Key points
- AI as a broad concept
- The place of machine learning
- The relationship with deep learning
Notes
- Environment: Paper exercise with AWS documentation and three hypothetical systems
- Consider how each system obtains its decision logic.
Foundation review
Where the three examples fit
An expert system can reason with rules written by people. A trained decision tree is machine learning, and an image classifier that learns representations with a multilayer network uses deep learning.
Training and deployment
A deployed tree can follow fixed branches and still be machine learning. The distinction is whether those branches were learned from training data.