Skip to content
Exit presentation

How AI Learns I: From Rules to Data

The big picture

The eight ideas, in order.

How AI Learns I — from a rule and a pile to one honest number
IdeaRemember
A rule or a pileA written rule and a pile are measured the same way. Close counts can hide different mistakes.
Keep some backA split makes the learn set and the kept-back set. Crates it never saw test it fairly.
Right out of how manyWalk the guesses and the real words together. Add one on each match.
Split it differentlyMove the split and the count moves. The count and the size of the test set are reported together.
Learning the examplesA nearest-example program is right about crates it holds. That is automatic. Only a new crate can surprise it.
Ask more than oneThree nearby examples can outvote one odd example. More examples do not always help.
The order you do it inSplit, then learn, then measure on the kept-back part. Measure on what it learned from and the count flatters.
Write the rule yourselfWalk the examples, carry the nearest so far, hand back its word.

You built a program and a count. The count says how often the program is right.

Next: Using AI Well & Safely I.

Jump to lesson

This track

  1. 0 Start here
  2. 1 A rule a person wrote, and a pile of examples
  3. 2 Keep some crates back
  4. 3 Right out of how many
  5. 4 Split it differently
  6. 5 Learning the examples instead of the pattern
  7. 6 Ask more than one
  8. 7 The order you do it in
  9. 8 Write the rule yourself
  10. 9 The big picture
  11. 10 Track quiz

Other tracks