Essay

Explain the relationship between training set size and dev-set error.

Question: In the context of plotting a learning curve, analyze why the dev-set error is generally expected to decrease as more examples are added to the training set. What does this relationship indicate about the model's ability to generalize?

Sample answer: As the training set size increases, the model is exposed to a broader and more representative variety of examples from the underlying data distribution. This helps the model learn more robust features and reduces overfitting to small idiosyncrasies in a small dataset. Consequently, the model's ability to generalize to new, unseen data improves, leading to a decrease in the dev-set error.

Key points:

  • More data exposes the model to a better representation of the distribution.
  • It reduces the model's tendency to overfit a small sample.
  • Improved generalization directly lowers the dev-set error.

Rubric: The response should correctly state the inverse relationship between training set size and dev-set error, explaining that more data improves generalization and reduces overfitting.

0

1

Updated 2026-05-27

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