Learn Before
Choose remedies for two teams facing different diagnosed errors.
Case context: Team A has diagnosed high avoidable bias in its neural network. Team B has diagnosed high variance in its system.
Question: Using only the simple remediation formula, what should each team do, and why?
Sample answer: Team A should increase the size of its neural network, potentially by adding layers or neurons, because its system has high avoidable bias. Team B should add data to its training set because its system has high variance.
Key points:
- Team A should increase model size.
- Adding layers or neurons is a valid example for Team A.
- Team B should add training data.
- Each action must be linked to the corresponding diagnosed issue.
Rubric: The response should assign a larger model to Team A and more training data to Team B, with each recommendation tied to the correct diagnosis.
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
Increasing Model Size to Reduce Avoidable Bias
Adding Training Data to Reduce Variance
Scaling Model Size and Training Data Has Practical Limits
Techniques for Reducing Avoidable Bias
Techniques for Reducing Variance
Which action does the simple formula recommend for high avoidable bias?
True or false: High variance calls for adding training data.
For high avoidable bias, increase the _____ of the model.
Match each bias/variance concept with its role in the simple formula.
Order the reasoning steps for applying the simple remediation formula.
Explain how the simple formula assigns different remedies to bias and variance.
Choose remedies for two teams facing different diagnosed errors.
State the two issue-to-action mappings in the formula.
Which pair correctly applies the formula to both diagnoses?
True or false: The formula gives the same remedy for bias and variance.