Learn Before
Effect of more training data on validation error
Question: How should dev-set error usually change if the same model is trained with a larger training set?
Sample answer: It usually decreases.
Key points:
- More training examples are added.
- Dev-set error tends to go down.
Rubric: Give full credit when the answer states that validation or dev error is expected to decline as the amount of training data increases.
0
1
Tags
Machine Learning
Deep Learning
Supervised Learning
Dive into Deep Learning @ D2L
Data Science
Machine Learning Strategy
Machine Learning Yearning @ DeepLearning.AI
Related
What trend should you expect in dev-set error as the training set gets larger?
True or False: Increasing the training set often reduces dev-set error when the model can benefit from more data.
More Training Data and Dev Error
Match each learning-curve term to its description.
Order the steps for creating a validation learning curve.
A team trains on 2,500 examples and sees 24% dev-set error. After adding 7,500 more training examples, what outcome should they expect?
True or False: Adding more training examples usually makes dev-set error worse.
What a Learning Curve Plots
Match each learning-curve pattern to its most likely interpretation.
Order the steps for deciding whether more training data is worth collecting based on a dev-set learning curve.
Why does validation error usually fall as the training set gets larger?
How a training-set increase should affect dev error
Effect of more training data on validation error