Averaging Errors from Several Small Training Samples
When a learning curve is too jagged to interpret clearly, you can draw several training samples of the same small size, train a separate model on each sample, measure each model's training and development error, and then average those errors before plotting the curve.
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
Why can a learning-curve value jump around when it is measured on a tiny random training sample?
A tiny random training sample can make the measured learning-curve error swing noticeably up or down.
A tiny sample with many unclear or incorrect labels is unusually _____.
Match each small-sample condition to its effect on a learning-curve estimate.
Put the checks in a sensible order for explaining a strange value on a tiny sample.
Why can a training curve wobble when the sample size is very small?
Explain a spike caused by a tiny imbalanced sample.
Why can learning-curve values be especially erratic for very small training samples?
Which situation makes a tiny random sample least likely to represent the full dataset?
A very small random sample is less likely to be misleading when the number of classes is large.
Proportionally Stratified Small Subsets Reduce Learning-Curve Noise
Learn After
Decide First Whether Learning Curves Need Smoothing
What should you do separately for each randomly drawn subset when estimating a learning curve?
Averaging Repeated Runs to Smooth a Curve
After training separate models on many random samples of the same size, you plot the _____ training error and validation error across runs.
What is the main benefit of averaging dev-set learning curves from many randomly chosen training subsets?
The averaging method creates several smaller training subsets by sampling with replacement from the original data.
How Many Models Should Be Trained?
Match each term in the resampling method to its correct description.
Order the steps for averaging learning curves over multiple random mini-samples.
When estimating a learning curve by averaging over random subsets, how many subsets are usually sampled?
True or False: In the averaging method, one model is fit to the union of all randomly sampled training subsets, and that single model’s error is then averaged.
Errors recorded before averaging learning curves
Match each problem with the part of the averaging method that addresses it.
Arrange the steps for using repeated subset experiments to clarify a noisy learning curve.
Why do repeated random subsamples make learning curves less noisy, and how is the procedure carried out?
Estimating Curve Values from Repeated Resamples
Resampling Rule and Curves for Reducing Variability