Learn Before
What should be improved after the verification test points to a scoring issue?
Question: In one to three sentences, explain where a developer should direct effort after the verification test shows that the scoring step is the problem.
Sample answer: The next step is to make the scoring model better at estimating the value for each input, not to spend time changing the search or optimization procedure.
Key points:
- Improve the score estimator or learning method
- Leave the optimization method alone for this issue
Rubric: Full credit for saying the focus should be on improving the learned score estimate; no credit for saying the optimization method should be the main target instead.
0
1
Tags
Data Science
Machine Learning
Deep Learning
Supervised Learning
Dive into Deep Learning @ D2L
Machine Learning Strategy
Machine Learning Yearning @ DeepLearning.AI
Related
In a product search ranking task, when does a scoring function problem occur?
True or False: If the correct label is not getting the highest score, the main problem is usually the optimizer.
A ranking rule is broken if the correct choice does not earn a _____ score than the wrong choice.
Match each score comparison result to the right diagnosis.
Put the diagnostic steps for a score-function failure in order.
Why a Diagnostic Check Needs a Strictly Larger Score
Diagnose a ranking failure with an optimization verification check
What should be improved after the verification test points to a scoring issue?
What should a team improve when the score ranks bad outputs ahead of good ones?
True or False: A system can have a scoring problem if the probabilities used to compute its score are estimated poorly.