When Development and Test Sets Reflect Different Populations
If the development set and test set come from different distributions, a better score on the development set does not reliably predict better test performance. That uncertainty makes it harder to know which changes are truly helping and harder to choose what to improve next.
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
Some Data Should Be Left Out of Training
When Development and Test Sets Reflect Different Populations
How Model Capacity Changes the Risk of Mixing Data Sources
One Predictor Can Work Across Multiple Data Sources
Choose evaluation data to match the real-world target
Mismatched Auxiliary Data Source
Building Dev and Test Sets Before Real Users Exist
Refreshing Evaluation Sets After a Product Launch
Using Public Web Images When No Better Future-Like Data Exists
Judging How Much to Invest in Dev and Test Sets
What should determine dev and test set selection?
True or False: You can assume the training set and test set always come from the same distribution.
Development and test sets should reflect the conditions you expect after deployment, not only the _____ available in your training pool.
Why can a simple random test split be a poor choice when the data you expect in the future is different from the data you have now?
You can usually assume the data used for training and the data used for testing come from the same distribution.
Design Dev and Test Sets for the Future
Match each concept about development and test sets to its description.
Order the steps for choosing development and test sets when future data differs from training data.
What should dev and test examples be designed to resemble?
A validation and test set must exactly match the training distribution in every project.
How should a test set be chosen when deployment data will differ?
Match each data scenario to the best dev/test set choice.
Order the reasoning steps for deciding whether a dev/test split is appropriate.
Why a Random 30 Percent Split Can Be Misleading When Future Data Will Differ
Dev and Test Splits for a Field-Photo Classifier
How to Choose Dev and Test Data When Future Data Will Differ
Learn After
What remains uncertain when the validation and test sets come from different distributions?
A change in development and test distributions can make it harder to choose which model issue to fix first.
Different dev and test distributions add extra uncertainty to model evaluation
Match each distribution issue to its downstream consequence.
Order the reasoning chain when the development set and test set come from different populations.
Why do different development and target data make model improvement harder to judge?
Explain why a stronger dev result may not justify the next engineering priority.
Why can a validation-set gain be less trustworthy when the validation and test sets come from different distributions?
Why does a dev/test distribution mismatch make it harder to choose fixes?
A higher score on a mismatched development set guarantees better test performance.