Practice free →
HomeGCP ML Engineercloudcomputingmletraining › Which interpretability requirement should an ML …

Which interpretability requirement should an ML Engineer balance against accuracy when CHOOSING a model architecture for a regulated financial workload — PMLE §3.1?

ACloud Memorystore
BAlways pick the deepest opaque neural net regardless of explainability needs
CCloud DNS
DPrefer interpretable architectures (linear / tree / GBDT) when regulators / stakeholders require feature-level explanations, accepting some accuracy trade-off
Answer & Solution
Correct answer: D. Prefer interpretable architectures (linear / tree / GBDT) when regulators / stakeholders require feature-level explanations, accepting some accuracy trade-off
Interpretability vs accuracy trade-off is a PMLE §3.1 design concern. The other options ignore the trade-off.
Solve this in the app — GCP ML Engineer practice & 24k+ MCQs →
Related questions