Skip to content
Exit presentation

How AI Learns I: From Rules to Data

Ask more than one

One odd example can pull the answer the wrong way. What if three examples vote?

Loading lab…

A new crate at 6 marked, 9 kilos.
Step 1 of 4
  1. A new crate at 6 marked, 9 kilos.
  2. Nearest one example: full price. Wrong.
  3. Nearest three examples: one full, two cut.
  4. The vote: cut wins, 2 to 1. Right.

The rule

Then the word with the most votes wins. So one odd example can be outvoted. The majority is the word most examples give. A tie means both words have equal votes.

Ask one, three or five nearby examples and let the majority answer. A split vote is reported as a tie. The count shows whether more examples helped.

At 12 marked and 12 kilos, the real word is cut price. One heavy full-price example is nearest, so asking one gives full. The three nearest are one full and two cut. Cut wins, 2 to 1. The vote is now right.

Two piles are tried. The whole pile is all 24 crates. The lightly marked half is 13 crates. Every crate in the half has 7 or fewer marked. Each is scored on the next morning's twenty.

Asking more nearby examples: two piles, two answers
PileAsk 1Ask 3Ask 5
Whole pile14 of 2015 of 2015 of 20
Lightly marked half13 of 2010 of 208 of 20

On the whole pile, asking three helps. On the lightly marked half, more voices hurt. Most examples there are full price. So a cut-price crate is outvoted more often.

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