Short Answer

Justify error attribution when the breed classifier output matches human prediction on a bad crop.

Question: Why is the cat detector, rather than the breed classifier, attributed with the error when the classifier outputs y=0 for a cropped image of a pile of rocks?

Sample answer: The cat detector is at fault because the breed classifier outputted a correct label (y=0) given its actual input of rocks, which matches how a human would classify that crop.

Key points:

  • The breed classifier predicted y=0, which is correct for the rocks it received.
  • A human shown the same input would also have predicted y=0.
  • The error belongs to the cat detector for cropping the wrong region.

Rubric: The answer should identify that the breed classifier is blameless because it classified its input correctly/reasonably, which matches human performance, meaning the fault lies with the cat detector's incorrect crop.

0

1

Updated 2026-05-26

Contributors are:

Who are from:

Tags

Machine Learning

Deep Learning

Supervised Learning

Dive into Deep Learning @ D2L

Data Science

Machine Learning Strategy

Related