When to prioritize improving one pipeline stage
Question: In a system made of several pipeline stages, when is it a strong choice to focus your improvement effort on one specific stage?
Sample answer: It is a strong choice to focus on a stage when its performance is much worse than the human baseline.
Key points:
- Compare the stage to a human-level reference.
- A large gap means that stage is a good candidate for improvement.
Rubric: The answer must say that the stage is far from human-level performance as the reason to prioritize improving it.
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
What is the best response when one stage in a multi-step system is far worse than expert performance?
True or False: If a subsystem performs much worse than expert-level performance on the same task, it is a strong candidate for targeted improvement.
If a pipeline stage is far from _____ performance, it is a strong candidate for improvement.
Match each model-status description to the right improvement priority.
Order the steps for using human-level comparisons to decide which subsystem to improve first.
Why should a subsystem far below expert-level performance usually receive top attention?
True or False: The best place to focus improvement is a module that already matches human performance closely.
When one module is far worse than human performance, you have a strong _____ to improve that module.
Match each model-performance scenario to the most appropriate takeaway.
Order the reasoning chain for choosing the weakest stage in a multi-step system to improve first.
Using a Human Benchmark to Choose Where to Improve a Model
Choosing the Bottleneck in a Fraud-Review Pipeline
When to prioritize improving one pipeline stage