AWS AI Practitioner - Practice Test 6 (Part 1)

Question 1

A healthcare company wants to use Amazon Bedrock to summarize medical records. They have a strict requirement that their data must never be used by AWS to improve the base models. What must they do to meet this requirement?

Accepted Answer

Nothing, Amazon Bedrock does not use customer data to train base models by default.

Answer

They must use Amazon SageMaker instead, as Bedrock does not offer this guarantee.

Answer

They must contact AWS Support to request an 'opt-out'.

Answer

They must use only 'Custom' models, as base models always use user data.

Question 2

Which machine learning technique is used in 'Amazon SageMaker Canvas' to automatically select the best algorithm and data preprocessing?

Accepted Answer

AutoML (Automated Machine Learning)

Answer

Reinforcement Learning

Answer

Unsupervised Clustering

Answer

Generative Adversarial Networks (GANs)

Question 3

What are TWO use cases where using 'Amazon Titan Embeddings' would be recommended? (Select TWO)

Accepted Answer

Answer

Audio-to-text transcription

Answer

Semantic document search

Answer

Generation of photorealistic images

Answer

Recommendation and personalization systems

Answer

Summarization of long texts

Question 4

Which Responsible AI metric helps identify if a model performs worse for a specific demographic group compared to another?

Accepted Answer

Accuracy Disparity

Answer

Faithfulness

Answer

Robustness

Answer

Latency

Question 5

A developer needs to build an application that summarizes legal documents. The full document is 50,000 words long. Which model feature is critical for this task?

Accepted Answer

Large Context Window

Answer

Audio support

Answer

Low latency

Answer

Image generation capability

Question 6

Which AWS service allows automatically detecting sensitive information (PII) in an Amazon S3 bucket and generating alerts?

Accepted Answer

Amazon Macie

Answer

AWS Shield

Answer

Amazon GuardDuty

Answer

Amazon Inspector

Question 7

Which 'Prompt Engineering' technique is useful when the model needs to reason about a problem but might benefit from seeing the correct answer before explaining it (reverse thinking)?

Accepted Answer

Generated Knowledge Prompting

Answer

Negative Prompting

Answer

Zero-shot

Answer

Chain-of-Thought

Question 8

Which AWS service is used to deploy custom inference containers if full OS access and complex network configuration are needed?

Accepted Answer

Amazon EC2 (or ECS/EKS on EC2)

Answer

Amazon SageMaker Serverless Inference

Answer

AWS Lambda

Answer

Amazon Bedrock

Question 9

What are TWO benefits of using 'Amazon SageMaker Role Manager'? (Select TWO)

Accepted Answer

Answer

Eliminates the need to use AWS IAM.

Answer

Automatically encrypts training data.

Answer

Allows public access to SageMaker notebooks.

Answer

Automatically generates least-privilege IAM policies based on predefined ML activities.

Answer

Provides an interface to define custom permissions for SageMaker users in minutes.

Question 10

Which AWS service is used for visual 'Defect Detection' on manufacturing lines using computer vision?

Accepted Answer

Amazon Lookout for Vision

Answer

Amazon Rekognition

Answer

Amazon Monitron

Answer

AWS Panorama