Question

You operate a deployed Vertex AI model that predicts loan default risk. The input features include several continuous numeric fields (income, loan amount) and categorical fields (employment type, region). You want to configure model monitoring to detect input feature drift and alert you when incoming serving data diverges significantly from the training baseline. Your ML lead asks which statistical measures Vertex AI Model Monitoring uses to quantify drift for each feature type. Which answer correctly maps the drift metric to the feature type?

Accepted Answer

Jensen-Shannon divergence for numerical features and L-infinity distance for categorical features

Answer

L-infinity distance for numerical features and chi-squared test for categorical features

Answer

Kolmogorov-Smirnov statistic for numerical features and cosine similarity for categorical features

Answer

Mean absolute error for numerical features and accuracy for categorical features