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 71–80 of 495

Q71

A company needs to run batch processing jobs regularly on large datasets. Which GCP solution should they choose?

  • A Cloud Run
  • B Cloud Dataflow
  • C Compute Engine
  • D App Engine
Explanation Cloud Dataflow is optimized for batch processing, unlike the others.
Q72

You are configuring a machine learning model with high variance. What is a recommended approach to improve the model?

  • A Increase training data size
  • B Decrease training data size
  • C Use linear regression
  • D Reduce feature set complexity
Explanation Increasing training data can help reduce overfitting due to high variance.
Q73

You are configuring a TensorFlow model in Vertex AI. What is the purpose of using a custom training pipeline?

  • A Handle data preprocessing automatically.
  • B Optimize hyperparameters effectively.
  • C Define and automate training workflows.
  • D Provide pre-trained models only.
Explanation Custom training pipelines automate the training workflow in Vertex AI, while other options do not represent the pipeline's primary purpose.
Q74

A company needs to classify images using a machine learning model. Which Google Cloud service can offer this functionality without managing infrastructure?

  • A Cloud Functions
  • B AI Platform Prediction
  • C Cloud Run
  • D BigQuery ML
Explanation AI Platform Prediction allows image classification without managing infrastructure, while other options do not specifically focus on machine learning classification.
Q75

What happens when you enable AutoML in Vertex AI for a dataset?

  • A Training begins immediately.
  • B It requires manual feature selection.
  • C Models will be trained automatically.
  • D No data exploration occurs.
Explanation Enabling AutoML allows for automatic training and model recommendation, while the other options imply incorrect manual requirements or lack of exploration.
Q76

Which Google Cloud service is best for managing APIs?

  • A Cloud Endpoints
  • B Cloud Functions
  • C Cloud Pub/Sub
  • D Cloud Storage
Explanation Cloud Endpoints is specifically designed to create, secure, and monitor APIs, while the others serve different purposes.
Q77

A company needs real-time processing of streaming data. Which Google Cloud solution should they use?

  • A Cloud Dataflow
  • B BigQuery
  • C Dataproc
  • D Cloud Storage
Explanation Cloud Dataflow is designed for real-time data processing, while the others focus on batch processing or storage.
Q78

You are configuring a machine learning model to predict customer churn. What happens when you set the threshold too low?

  • A High precision, low recall
  • B High recall, low precision
  • C Balanced accuracy
  • D High false negative rate
Explanation Setting a low threshold increases recall (capturing more positives) but decreases precision (more false positives).
Q79

Which Google Cloud service is used for developing machine learning models using pre-trained APIs?

  • A Cloud AI Platform
  • B BigQuery
  • C Cloud Functions
  • D Cloud Storage
Explanation Cloud AI Platform offers tools for model development; BigQuery handles data analysis, Cloud Functions is for serverless computing, and Cloud Storage is for file storage.
Q80

A company needs to create a data pipeline that processes real-time events. Which Google Cloud service should they use?

  • A Cloud Bigtable
  • B Cloud Pub/Sub
  • C Cloud Spanner
  • D Cloud Functions
Explanation Cloud Pub/Sub supports real-time messaging and streaming; Cloud Bigtable is a NoSQL datastore, Cloud Spanner is a relational database, and Cloud Functions runs event-driven code, but not as a full pipeline solution.