After training the shoe sorter, what can Nora try next?

• Delete all examples before learning from them
• Make a guess about a new photo
• Stop using the learned pattern on any photos

For a new picture sorter, what comes before training it on cat and dog groups?

• Give it labeled cat and dog examples
• Ask for guesses without giving any examples
• Delete the pictures before it can use them

In machine learning, a person must write a separate sorting rule for every possible picture.

Circle one:   True   False

What is the aim of training the sneaker-and-boot picture sorter?

• Learn features that help separate the two shoe groups
• Store only filenames and ignore both groups of shoes
• Count photos but never use features in either shoe group

Nora checks a new photo whose known label is 'sneaker'. Her trained model guesses 'boot'. What does this check show?

• The model correctly labeled this picture as a sneaker
• Every future shoe prediction must also be a mistake
• The model made a mistake on this picture

In this made-up test, both models tried 10 new photos. The chart shows correct guesses. How many more did the 200-photo model get right?

Answer: ______________

A shoe sorter labels rainy photos as 'boot', even when they show sneakers. Which training problem could explain its use of rain as a clue?

• Both shoe groups had equally varied weather in every photo
• The labels gave no link between rain and either shoe group
• Rain appeared mainly in the boot training photos

Two cat-and-dog models use the same dog photos. Amir adds 500 copies of one cat photo. Priya adds 100 photos of different cats in different places. Which model has more useful cat variety?

• Amir's, since repeating one cat covers every type of cat
• Priya's, since her cat photos have more variety
• Both equally, since photo variety never affects learning