Remember: A machine learns patterns from many examples, but it can be mistaken, so people check its work and people decide.
Read and answer
Read the story, then answer. In the chart, Varied means lots of different examples.
Cat or dog?Mr. Reyes brings a pet sorter that looks at a photo and says cat or dog. Mina tries writing rules, like pointy ears mean cat, but some dogs have pointy ears, so the rules break. Mr. Reyes has a different idea. A hundred photos, each labelled cat or dog, go into the machine, and it notices patterns. That is machine learning. Teo holds up a brand-new photo, and the sorter says dog, and it is right. Then Zoe holds up her fluffy dog snoozing on a sofa, and the sorter says cat! Nearly every cat photo had a sofa in it, so the machine learned a pattern nobody meant. It does not understand pets, and only finds patterns in what it is shown. So Mina adds many different photos, and the machine learns again.
1
How did Mr. Reyes teach the pet sorter?
📏By writing a rule for every ear and tail🪄By magic🖼️By showing it many labelled photos
2
Each sorter tried ten new photos. Which one got the most right?
🌱Few examples🌿Lots of examples🌳Varied examples
3
Each sorter tried ten new photos. How many did the Lots sorter get right?
579
4
A machine finds patterns. It does not understand pets.
TrueFalse
5
A machine is never mistaken.
TrueFalse
6
Why did the sorter say cat for Zoe's dog?
It understood the dog was a catZoe's dog has pointy earsNearly every cat photo had a sofa, so it learned sofa means cat
Match, order and fix
7
Machine learning finds patterns in many .
rulesexamplessofas
8
Join each beginning to its ending.
1A learning machine
2Varied examples
3For things that matter
4Without examples
Apeople check and decide
Bteach more than copies do
Ccan be sure and mistaken
Da learning machine finds no pattern
9
Put the story in order.
Mina adds many different photos
Mina writes rules, and they break
The sorter says cat for Zoe's dog on a sofa
Mr. Reyes gives the machine a hundred labelled photos
10
Zoe says: the sorter said cat so confidently that it must be right. What is the mix-up to fix?
A machine can be sure and mistaken, so people checkConfident machines are always rightDogs are cats
11
How many more photos did the Varied sorter get right than the Few sorter?
12
Write one way people can check a sorter's work.
13
Draw three different photos you would add to teach the sorter about dogs.
Scan the code at the top to mark this sheet on screenTake a photo of your answers or type them. It is marked for you automatically, and your score is kept.
Answer key · Set A
The pet sorter
1By showing it many labelled photos Rules kept breaking, so he gave the machine a hundred photos labelled cat or dog to learn patterns from.
2Varied examples The Varied bar is the tallest, at nine out of ten.
37 The Lots bar reaches seven.
4True It only knows what its examples showed it.
5False A machine can be sure and mistaken, so people check its work.
6Nearly every cat photo had a sofa, so it learned sofa means cat The machine learned a pattern nobody meant, because the dog was on a sofa.
7examples A learning machine needs examples to find patterns.
81-C, 2-B, 3-A, 4-D These are the lines the story gives: a machine can be sure and mistaken, varied examples teach more, people decide what matters, and examples come first.
94, 1, 3, 2 The rules break first, then examples, then the sofa mistake, and then different photos to fix it.
10A machine can be sure and mistaken, so people check A machine can be sure and mistaken. People check its work.
114 Varied got 9 and Few got 5, and 9 - 5 = 4.
12Your own words: a grown-up reads it.
Mark each answer, count the right ones and write the score on the sheet.