Google Cloud

Google Cloud Certified – Professional Machine Learning Engineer

PR000269
Trending

Validate your skills as a Professional Machine Learning Engineer with exam code PR000269 in Google Cloud.

495 questions 0 views Free
Start Mock Test Timed · Full-length · Scored

Questions 351–360 of 495

Q351

You are configuring a Vertex AI pipeline. What happens if a component fails?

  • A The pipeline retries automatically.
  • B The pipeline immediately stops.
  • C Only successful components proceed.
  • D The pipeline sends an alert.
Explanation If a component in a Vertex AI pipeline fails, the pipeline stops immediately unless configured for retries.
Q352

Which service is best for deploying machine learning models in production on Google Cloud?

  • A AI Platform
  • B Cloud Functions
  • C BigQuery
  • D Cloud Storage
Explanation AI Platform is specifically designed for deploying machine learning models, while the others serve different purposes.
Q353

A company needs to process a substantial amount of real-time data with complex event processing. Which Google Cloud service should they use?

  • A Cloud Pub/Sub
  • B Cloud Dataflow
  • C BigQuery
  • D Cloud Functions
Explanation Cloud Dataflow handles real-time data processing, whereas the other options have different capabilities.
Q354

You are configuring data retention policies in BigQuery. What happens when you set a table's expiration time to 0?

  • A Data never expires
  • B Data expires immediately
  • C Data expires in 24 hours
  • D Data expiration is disabled
Explanation Setting expiration time to 0 means the data will never expire; other options misrepresent how expiration works.
Q355

Which service would you use for training large-scale machine learning models on Google Cloud?

  • A AI Platform
  • B Cloud Functions
  • C BigQuery
  • D Cloud Storage
Explanation AI Platform is designed for training ML models, while the others serve different purposes.
Q356

A company needs to deploy a machine learning model but wants to ensure minimal downtime during updates. Which approach should they take?

  • A Rolling update
  • B Blue-green deployment
  • C Canary deployment
  • D N/A
Explanation Blue-green deployment allows seamless model updates without downtime, unlike the other methods.
Q357

What happens when you set a Cloud Function to trigger on a Pub/Sub message, but the processing takes longer than the function timeout?

  • A Message is successfully processed
  • B Function retries immediately
  • C Message is sent to a dead-letter queue
  • D Function fails and message is lost
Explanation If the processing exceeds timeout, the function fails, leading to message loss if not handled correctly.
Q358

Which service provides automated ML model tuning?

  • A AI Platform Hyperparameter Tuning
  • B BigQuery Machine Learning
  • C Dataflow for Training
  • D Cloud AutoML Tables
Explanation AI Platform Hyperparameter Tuning specifically automates model tuning; other options serve different purposes.
Q359

A company needs to securely share machine learning data. Which service is best suited for this?

  • A Google Drive
  • B Cloud Storage with IAM
  • C BigQuery Dataset
  • D Cloud Pub/Sub
Explanation Cloud Storage with IAM allows fine-grained access control, unlike the others which lack secure sharing features.
Q360

What happens when a TensorFlow model has an excessively high learning rate?

  • A Model converges too quickly
  • B Model diverges and loses accuracy
  • C Model overfits the training data
  • D No effect, train runs normally
Explanation A high learning rate can cause divergence, while other options inaccurately describe learning rate impacts.