You are part of your organization's ML engineering team and notice that the accuracy of a model that was recently deployed into production is deteriorating. What is the best first step to address…

The correct answer is B. Conduct champion/challenger testing. Champion/challenger testing allows comparison between the current model and alternatives to identify if a replacement or adjustment is necessary, addressing accuracy deterioration

AI Governance Program

Question

You are part of your organization’s ML engineering team and notice that the accuracy of a model that was recently deployed into production is deteriorating. What is the best first step to address this?

Options

AReplace the model with a previous version.
BConduct champion/challenger testing.
CPerform an audit of the model.
DRun red-teaming exercises.

How the community answered

(32 responses)

A
6% (2)
B
72% (23)
C
16% (5)
D
6% (2)

Explanation

Champion/challenger testing allows comparison between the current model and alternatives to identify if a replacement or adjustment is necessary, addressing accuracy deterioration

Topics

#model performance degradation#champion-challenger testing#production monitoring#model retraining

Community Discussion

No community discussion yet for this question.

Full AIGP Practice