You work for an online retail company that is creating a visual search engine. You have set up an end-to-end ML pipeline on Google Cloud to classify whether an image contains your company's product. E

Question

Options

• AKeep the original test dataset unchanged even if newer products are incorporated into retraining.
• BExtend your test dataset with images of the newer products when they are introduced to
• CReplace your test dataset with images of the newer products when they are introduced to
• DUpdate your test dataset with images of the newer products when your evaluation metrics drop

• B
  84% (16)
• C
  5% (1)
• D
  11% (2)

Explanation

When new products are introduced, the test dataset must be extended (augmented) with images of those new products - not replaced (C), which would lose coverage of existing products, causing the model to appear accurate on new items while silently degrading on existing ones. Keeping the test set completely unchanged (A) means the evaluation service never tests the model's ability to classify new products, making the evaluation misleading over time. Waiting until metrics drop to update (D) is reactive - by the time metrics fall, the model is already underperforming in production. Extending the test set proactively (B) ensures comprehensive, continuous evaluation across both old and new product categories as the catalog evolves.

No community discussion yet for this question.