The big picture
Here are the five ideas, in order.
| Idea | Remember |
|---|---|
| Both at once | A machine can hold a written rule and a learned part. |
| Training | Labelled examples carry answers. Showing many of them is one way to train. The rule is learned from the pile. When the pile changes, the rule changes. |
| A decision tree | A chain of yes/no questions. Who chose them decides written or learned. |
| What it learned from | Pictures, words or tables point to vision, NLP or data science. |
| Trace the mistake | The path is the questions asked. A wrong answer is a missing question. |
Ask what a machine was shown. Then ask who chose its rule.