Review Questions

Q1 - D ✅

Q2 - C ✅

Q3 - ?? C ✅

• no AutoML for matrix factorisation today
• recommendations is a special type of classification

Q4 - A ✅

Q5 - C ✅

Q6 - C ✅

Q7 - C ✅

Q8 - C ✅

Q9 - D ✅

Q10 - B ❌ D

• A (Export to GCS, then automate transfer to BigQuery):

  More complex, involves extra steps and automation overhead.

• B (Move your model to BigQuery ML):

  Not practical if your team extensively uses Vertex AI AutoML Tables and pipelines; migrating models is costly and unnecessary.

• C (Batch prediction → GCS → BigQuery):

  Similar to A, an extra step of moving data, which adds complexity and latency.

Q11 - C ✅

Q12 - A ✅

Q13 - A ✅