← Back to course
AI Explorers K-2 / Week 06 / Explorer Quest
6/6
Week 06 · Learn From Examples

Explorer Quest

Explorer Quest: Blorp or not?
// Training data
⏱ about 15 min

Explorer Quest: Blorp or not?

Raven invents a blorp.

Blorps have three legs.

Rocket guesses fast. Oops!

Nova hints, "Look at more examples."

  1. We learn from many examples.
  2. Computers can learn from data.
  3. Training data are the examples.
  4. People usually give the examples and labels.
  5. Computers find patterns. They do not feel.
◇ BLORP OR NOT?
Your grown-up picks a secret blorp rule.
They draw six cards: blorps and not-blorps.
They show each card and say its label.
Then they draw a new creature.
You guess: blorp or not? Then swap!
For a grown-up
Ask: what did you notice about blorps?
Ask: did more examples help?
Say: computers need a huge number of examples.
Next week explores examples that are one-sided.
WEEK CHECK
  • Read the question.
  • Tap your answer.
The blorp cards were like...
More examples help you...
QUEST CHECK
  • Read the question.
  • Tap your answer.
Rocket guessed too fast. What helps?
Who said "blorp" or "not"?
Is a blorp real?
Draw a blorp and a not-blorp.

You taught and learned like a pro!

Next week: more examples, better guesses.

← Friday