What is the underlying consequence of only correcting labels on misclassified dev examples?
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 does label-correction bias tend to creep into dev sets during error analysis?
True or False: Checking only misclassified dev examples is easier because there are far fewer of them to review.
With 1,000 dev examples and 98.0% accuracy, it is easier to examine the _____ misclassified examples than all 980 correct ones.
Match each dev set quantity to its description for a 1,000-example set with 98.0% accuracy.
Order the reasoning that explains how practical convenience leads to label-correction bias.
Analyze why practical convenience, rather than a deliberate methodological choice, drives label-correction bias in dev sets.
Diagnose why a team's practice of only reviewing misclassified dev examples introduces bias.
In one to three sentences, explain why reviewing only misclassified dev examples is more practical than reviewing all examples.
What is the underlying consequence of only correcting labels on misclassified dev examples?
True or False: The 980 correctly classified examples are guaranteed to have accurate labels since they were not flagged for correction.