What observation must justify averaging learning curves over multiple subsets?
Question: Answer in one to three sentences: What should you observe before deciding that averaging over multiple subsets is worth using?
Sample answer: You should first plot the learning curves and observe that they are too noisy to reveal the underlying trends. Without that finding, averaging over multiple subsets is not worth using.
Key points:
- Plot the learning curves first.
- Confirm that noise obscures the underlying trends.
Rubric: Credit answers that identify both prior plotting and noise severe enough to obscure the underlying trends.
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
When should you consider using techniques that reduce noise in learning curves?
Noise-reduction techniques should be applied before inspecting the initial learning curves.
Use noise-reduction techniques only if learning curves are too _____ to reveal underlying trends.
Match each learning-curve condition to its implication for noise-reduction techniques.
Order the decision process for using learning-curve noise-reduction techniques.
Explain why noise-reduction techniques should follow, rather than precede, an initial learning-curve plot.
Decide whether a team should add noise reduction to already readable learning curves.
What observation must justify averaging learning curves over multiple subsets?
Which dataset most strongly suggests that learning-curve noise reduction is unnecessary?
A large, not-very-skewed dataset usually removes the need for these noise-reduction techniques.