Learning from Examples
This video presents the same text shown beside it, spoken and on screen. It adds nothing the text does not say.
State
Supervised learning turns training examples — inputs paired with correct outputs — into a predictor for new cases shown only their inputs.
Show
The inputs are features, the output is the target, and the examples are the only teacher present.
Watch for
Whatever the predictor comes to know, it must extract from the pile — nobody supplies rules or theory.