Question 7
Content Domain 3: ModelingA machine learning team receives new production data and wants the existing model to reflect the latest patterns without rebuilding the entire training process from scratch. Which action best matches updating the model rather than retraining it?
Correct answer: A
Explanation
Updating applies changes to an existing model so it reflects new information, while retraining rebuilds the model through a new training process using data again. — Source material: "Update and retrain models." Key Terms: update, retrain.
Why each option is right or wrong
A. Modify the existing model so it incorporates new information without starting the full training process over
The source distinguishes "update" from "retrain." In this scenario, the team wants the current model to reflect new production patterns without rebuilding the entire training process, which aligns with updating the existing model rather than running training again from the beginning.
B. Run the complete training process again on data to produce a newly trained version of the model
Retraining means running training again to rebuild the model, not modifying the current one in place.
C. Leave the model unchanged and only monitor whether its predictions remain stable over time
Updating changes the model; monitoring alone does not alter the model at all.
D. Replace the model with a different algorithm before using any new data in the workflow
Updating concerns changing the existing model, not switching to a different model type first.