Short Answer

Why choose a replacement metric instead of hand-picking models?

Question: When a team no longer trusts its current evaluation metric, why is it better to define a new metric and let that become the team goal rather than manually choosing a model each time?

Sample answer: A new metric gives the team a stable objective to optimize directly. That is better than repeatedly hand-selecting models, which is slow, hard to scale, and keeps decision-making tied to one-off judgment instead of a shared target.

Key points:

  • Creates a concrete goal the whole team can work toward
  • Reduces ad hoc manual model selection
  • Supports more consistent, scalable progress

Rubric: Answer should explain that a replacement metric provides a clear team objective and avoids the inefficiency of manually selecting classifiers.

0

1

Updated 2026-08-12

Contributors are:

Who are from:

Tags

Machine Learning

Deep Learning

Machine Learning Strategy

Supervised Learning

Dive into Deep Learning @ D2L

Data Science

Machine Learning Yearning @ DeepLearning.AI