#generalization-error

1 question from Learning Theory

Timed test

1 question · 1 min · score and explanations at the end

Q 1 / 1Score 0 / 0

Shuffle questions

Question 1: Single choice

mediumAI-checkedLearning Theory

A classifier reaches zero training error. Under which condition can its generalization (population) error also be zero?

Labels are a noise-free function of the input, and the model learned that function1
The training set is very large, whatever the amount of label noise in it2
The model class is flexible enough to fit every labelling of the training set3
The training error is zero, and a zero training error already implies zero test error4

Keys: 1–9 select · Enter check / next