Short Answer

Why can learning-curve values be especially erratic for very small training samples?

Question: Answer in one to three sentences.

Sample answer: With very small samples, random differences in which examples are selected matter a lot. A sample that happens to be unusually clean and representative can produce a better point, while a sample with difficult or mislabeled examples can produce a worse one, so the curve can swing noticeably.

Key points:

  • Small samples are highly sensitive to random selection
  • Better samples can lower the measured error
  • Worse samples can raise the measured error
  • This sensitivity causes visible fluctuation at small training sizes

Rubric: The answer should explain that unusually strong or weak small subsets change the measured point and make the curve noisy.

0

1

Updated 2026-08-12

Contributors are:

Who are from:

Tags

Machine Learning

Deep Learning

Supervised Learning

Dive into Deep Learning @ D2L

Data Science

Machine Learning Strategy

Machine Learning Yearning @ DeepLearning.AI