Case Study

Case Study: Setting a desired error rate using long-term application experience.

Case context: A team has worked on an important speech application for several years. They are now trying to decide what desired error rate to target for the next quarter's model.

Question: Based on the source concept, what should the team do to set their desired error rate, and what kind of intuition should they draw on?

Sample answer: The team should draw on their long-term experience with the application to form an intuition about how much more progress is reasonably achievable in the next quarter. Since they have worked on this important application for a long time, they are well positioned to judge realistic improvement, and this intuition should directly inform the desired error rate they set for the next quarter.

Key points:

  • Team has worked on the application for a long time
  • This experience gives intuition about reasonable progress next quarter/year
  • That intuition should inform the desired error rate
  • Distinct from product-experience intuition, which is about user experience

Rubric: Full credit for identifying that long-term experience with the application should be used to judge reasonably achievable progress, and that this judgment informs the desired error rate.

0

1

Updated 2026-07-10

Contributors are:

Who are from:

Tags

Machine Learning

Deep Learning

Supervised Learning

Dive into Deep Learning @ D2L

Data Science

Machine Learning Strategy

Machine Learning Yearning @ DeepLearning.AI