Question 1

Your team has developed a machine learning model for your customer. The test results indicate very strong predictive capability. The model is then deployed in production. Evaluation of the predictions in production show that they are off by a pronounced margin. What is the issue and how can you solve for it?

Options :

A : The model is under fitted. Train with less data.

B : The model is over fitted. Add more features to the model to fix it.

C : The model is fine since the test results are good. Fix the production of incoming data.

D : The model is overfitted. Train with more data.

Answer: D

Question 2

Domain: Google Cloud products and services

Which feature of Cloud Key Management Service helps prevent accidental delete?

Options :

A : Built-in 12-hour delay for key material destruction

B : Built-in 24-hour delay for key material destruction

C : Ability to create external keys with EKM

D : Ability of at-will key rotation

Answer: B

Question 3

When creating machine learning models, a key initial step is to identify the type of model required. One of these is the classification model. Which of these statements define a classification model?

Options :

A : A type of machine learning model for distinguishing among two or more discrete values. E.g. 'book', 'car'.

B : A type of machine learning model is a meta-model maker, which classifies algo-rithms based on the quality of their output.

C : A type of machine learning model that outputs continuous (typically, floating-point) values. E.g. the predicted price of the house is $120,000.

D : A type of classic model approach that is less used today and which has been re-placed by the regression model.

Answer: A

Question 4

Domain: Google Cloud products and services

Which of the below listed Pub/Sub, Cloud Tasks feature is incorrect?

Options :

A : Implicit invocation is supported by Pub/Sub

B : Explicit invocation is supported by Cloud Tasks

C : Cloud Tasks are ideal for scenarios where task producers require no control over remote procedure calls.

D : Pub/Sub is an ideal choice for general event data ingestion.

Answer: C

Question 5

An organization runs their application on a virtual machine, but every time they want to edit specific features, they have to bring the system offline to update the applica-tion. What would be a more appropriate solution for their app?

Options :

A : GPUs

B : Containers

C : Hypervisors

D : Solid State Disk

Answer: B