Select Two (Multi-Select) Questions

Increase the max token limit to generate longer code.

Why wrong: Longer code does not improve security.

Set the temperature to 0 for deterministic output.

Why wrong: Deterministic output does not ensure security.

Use a larger model for better accuracy.

Why wrong: Larger models do not guarantee security.

Precision

Why wrong: Precision is for classification.

Accuracy

Why wrong: Accuracy is for classification tasks, not text generation.

F1 score

Why wrong: F1 score is for classification.

One-hot encoding

Why wrong: One-hot encoding is a categorical encoding technique.

Principal Component Analysis (PCA)

Why wrong: PCA is a dimensionality reduction technique, not a scaling method.

Binning

Why wrong: Binning is discretization, not scaling.

Amazon DynamoDB Accelerator (DAX)

Why wrong: DAX is a caching service for DynamoDB, not relevant to security compliance.

AWS WAF

Why wrong: WAF protects web applications, but the primary threat protection for ML pipelines is more comprehensive via GuardDuty.

Amazon GuardDuty

Why wrong: GuardDuty is a threat detection service for AWS accounts.

AWS WAF

Why wrong: WAF protects web applications from common exploits, not image content.

Amazon Comprehend

Why wrong: Comprehend is for text analysis, not images.

Use a high temperature parameter to increase creativity.

Why wrong: High temperature makes output more random and less focused.

Use a low frequency penalty to reduce repetition.

Why wrong: Frequency penalty addresses repetition, not summary quality.

Use a longer context length by increasing the max tokens parameter.

Why wrong: Increasing max tokens does not help if the document exceeds the model's context window.

Use a smaller model

Why wrong: Model size does not directly correlate with harmfulness.

Fine-tune the model on a biased dataset

Why wrong: Fine-tuning on biased data would likely increase bias.

Disable all logging

Why wrong: Disabling logging reduces accountability.

Bedrock Agents

Why wrong: Wrong: Bedrock Agents help build autonomous agents but do not inherently provide grounding or content filtering.

Amazon Comprehend

Why wrong: Wrong: Amazon Comprehend is for NLP insights, not for grounding or filtering.

Amazon SageMaker JumpStart

Why wrong: Wrong: SageMaker JumpStart provides pre-built models but not directly for grounding or filtering.

Increase model size to improve performance.

Why wrong: Model size should be chosen based on task; larger models may overfit or be costly.

Deploy the model to production immediately after fine-tuning.

Why wrong: Models should be evaluated and tested before production deployment.

Use larger instance types with more GPUs to speed up inference.

Why wrong: Larger instances are more expensive; they may reduce latency but not cost.

Store the model parameters in FP32 to maintain accuracy during inference.

Why wrong: FP32 uses more memory and compute, increasing cost without significant accuracy benefit.

Increase the number of training epochs to achieve higher accuracy.

Why wrong: More epochs increase training cost and do not reduce inference cost.

It can only be used for text data.

Why wrong: It supports image, video, and text.

It can only use a public workforce from Amazon Mechanical Turk.

Why wrong: It also supports private workforces (Vendors and your own employees).

Collect as much data as possible without quality checks

Why wrong: Data quality is more important than quantity.

Ensure the development team is homogeneous to avoid conflicts

Why wrong: Homogeneous teams increase bias risk.

Use the most complex model available for maximum accuracy

Why wrong: Complexity can reduce explainability.

Create an IAM role for Rekognition to assume and attach it to the Rekognition service

Why wrong: Rekognition cannot assume a role; the calling service must assume the role.

Configure the S3 bucket with a bucket ACL that grants read access to the Rekognition service

Why wrong: Bucket ACLs are not the recommended way; Rekognition uses IAM or bucket policies.

Attach a resource-based policy to the Rekognition collection allowing S3 access

Why wrong: Rekognition collections do not have resource-based policies.

AWS Glue

Why wrong: Glue is a managed ETL service, independent of SageMaker.

Amazon Athena

Why wrong: Athena is a serverless query service for S3, not part of SageMaker.

Create an Amazon ECR repository for the inference container

Why wrong: SageMaker provides a pre-built TensorFlow container; a custom container is not required.

Submit a training job to SageMaker

Why wrong: The model is already trained; no need for a new training job.

Convert the model to ONNX format

Why wrong: TensorFlow models can be deployed directly without conversion.

Audio feature extraction

Why wrong: Audio processing is not a built-in transform in Data Wrangler.

Image resizing and normalization

Why wrong: Image transformations are not natively supported; Data Wrangler focuses on tabular and text data.

Always split data 80/20 for training and testing

Why wrong: The split ratio depends on dataset size; 80/20 is not always optimal.

One-hot encoding for ordinal categories

Why wrong: Ordinal categories should use ordinal encoding; one-hot encoding ignores order.

Drop duplicate records only if they are manual entry errors

Why wrong: Duplicates should generally be removed to avoid bias, not only manual errors.

Add more irrelevant features

Why wrong: Irrelevant features introduce noise and can lead to overfitting.

Increase model complexity

Why wrong: More complex models are more prone to overfitting.

Reduce the amount of training data

Why wrong: Less training data makes overfitting more likely.

Use SageMaker Feature Store offline store with record identifier and event time.

Why wrong: This helps with time-based queries but does not inherently log definition changes.

Implement feature store online store with TTL to automatically expire data.

Why wrong: TTL manages data lifecycle, not immutability or change logging.

Use AWS Config to track changes to Feature Store resources.

Why wrong: Config tracks resource configurations, not specific feature definition changes.

Amazon Macie

Why wrong: Macie discovers sensitive data in S3, not real-time inference data.

AWS CloudTrail

Why wrong: CloudTrail logs API calls, not inference data patterns.

Amazon CodeGuru Security

Why wrong: CodeGuru reviews code for vulnerabilities, not runtime data.