HOTSPOT
-
A hospital wants to predict patient outcomes for the coming year. An ML engineer must improve several existing ML models that currently perform poorly.
Select the correct regularization method from the following list to improve each model. Select each regularization method one time, more than one time, or not at all. (Choose three.)
• L1 regularization
• L2 regularization
• Early stopping
Explanation: L2 (Ridge) shrinks coefficients without driving them to zero. L1 (Lasso) performs feature selection by setting irrelevant coefficients to zero, making it suitable for eliminating unnecessary polynomial terms and redundant correlated predictors.
Question 52
HOTSPOT
-
A company develops an ML model to classify products. The model uses textual data and visual data to classify the products into a hierarchical taxonomy. An ML engineer must use specific strategies to enhance the model’s accuracy and handle class imbalances.
Select the correct strategy from the following list for each use case. Select each strategy one time. (Choose five.)
• Categorical cross-entropy
• Fallback metric
• Hierarchical loss
• General oversampling
• Synthetic Minority Oversampling Technique (SMOTE) for text
Show Answer
Correct Answer: Evaluate predictions with tolerance for broader category matches → Fallback metric
Balance the training data by increasing the presence of minority categories → General oversampling
Enhance the model’s ability to recognize nested structures within product categories → Hierarchical loss
Enhance textual data representation for underrepresented categories → Synthetic Minority Oversampling Technique (SMOTE) for text
Explanation: Fallback metrics can give partial credit for broader-category matches. General oversampling increases minority-class frequency. Hierarchical loss incorporates taxonomy structure into training. SMOTE for text synthesizes minority text samples to improve representation.
Question 53
An ML engineer is using Amazon SageMaker Canvas to build a custom ML model from an imported dataset. The ML engineer wants the model to make continuous numeric predictions based on 10 years of data.
Which metric should the ML engineer use to evaluate the model’s performance?
A. Accuracy
B. InferenceLatency
C. Area Under the ROC Curve (AUC)
D. Root mean square error (RMSE)
Show Answer
Correct Answer: D
Explanation: The task is to make continuous numeric predictions, which is a regression problem. RMSE is a standard evaluation metric for regression because it measures the magnitude of prediction errors and penalizes larger errors. Accuracy and AUC are classification metrics, while inference latency measures performance speed rather than predictive quality.
Question 54
A company is preparing data to train a new ML model on Amazon SageMaker AI. The data has not been used before for ML training. The data includes duplicates and is missing some values.
The company needs to increase the data quality and detect any statistical bias in the data.
Which solution will meet these requirements?
A. Use SageMaker Clarify to create data quality rules. Use SageMaker Model Monitor to detect bias.
B. Use SageMaker Data Wrangler to create data quality rules. Use SageMaker Clarify to detect bias.
C. Use SageMaker Debugger to create data quality rules. Use SageMaker Model Monitor to detect bias.
D. Use SageMaker Model Monitor to create data quality rules. Use SageMaker Clarify to detect bias.
Show Answer
Correct Answer: B
Explanation: SageMaker Data Wrangler is designed for data preparation, profiling, and improving data quality, including identifying issues such as missing values and duplicates and applying data quality rules or transformations. SageMaker Clarify is the AWS service for detecting statistical bias in datasets before training and for model explainability. Model Monitor is primarily for monitoring data and model quality in production, and Debugger is for diagnosing training issues, not data quality preparation.
Question 55
A recommendation model uses ML and calls an Amazon SageMaker AI endpoint to get recommendations. An ML engineer must ensure that the model stays available during an expected increase in user traffic.
Which solution will meet these requirements?
A. Configure auto scaling on the SageMaker AI endpoint.
B. Create a new SageMaker AI endpoint. Deploy the model to the new endpoint.
C. Use SageMaker Neo to optimize the model for inference.
D. Attach an Auto Scaling group to the SageMaker AI endpoint.
Show Answer
Correct Answer: A
Explanation: Amazon SageMaker endpoints support Application Auto Scaling to automatically adjust the number of inference instances based on traffic and metrics, helping maintain availability during increased request volume. Creating another endpoint alone does not scale existing traffic, SageMaker Neo optimizes model compilation rather than capacity, and EC2 Auto Scaling groups are not attached directly to SageMaker endpoints.
Question 56
An ML engineer is using AWS Glue to transform proprietary data from a third-party vendor to a format that the ML engineer intends to use with the Amazon SageMaker DeepAR forecasting algorithm. The data includes several similar time series data files that the ML engineer must convert to the appropriate format. The ML engineer must compress the files to optimize storage costs.
Which solution will meet these requirements?
A. Use Snappy to convert the files to RecordIO-Protobuf and to compress the files.
B. Use XZ to convert the files to RecordIO-Protobuf and to compress the files.
C. Use XZ to convert the files to Apache Parquet format and to compress the files.
D. Use gzip to convert the files to Apache Parquet and to compress the files.
Show Answer
Correct Answer: D
Explanation: Amazon SageMaker DeepAR supports JSON Lines and Apache Parquet input formats for training, not RecordIO-Protobuf. AWS Glue can transform the time series data into Apache Parquet, and gzip is a standard supported compression option that reduces storage costs and is compatible with DeepAR.
Question 57
A company needs to ingest data from data sources into Amazon SageMaker Data Wrangler. The data sources are Amazon S3, Amazon Redshift, and Snowflake. The ingested data must always be up to date with the latest changes in the source systems.
Which solution will meet these requirements?
A. Use direct connections to import data from the data sources into Data Wrangler.
B. Use cataloged connections to import data from the data sources into Data Wrangler.
C. Use AWS Glue to extract data from the data sources. Use AWS Glue also to import the data directly into Data Wrangler.
D. Use AWS Lambda to extract data from the data sources. Use Lambda also to import the data directly into Data Wrangler.
Show Answer
Correct Answer: A
Explanation: Amazon SageMaker Data Wrangler supports native direct connections to Amazon S3, Amazon Redshift, and Snowflake. These connections read directly from the source when data is imported or refreshed, ensuring access to the latest available data. Cataloged connections rely on metadata catalogs rather than being the appropriate choice for always-current source access, and neither AWS Glue nor AWS Lambda imports data directly into Data Wrangler in the manner described.
Question 58
A logistics company has installed in-vehicle cameras for basic monitoring of its drivers. The company wants to improve driver safety by identifying distractions that could lead to accidents.
Which solution will meet this requirement with the LEAST operational effort?
A. Use Amazon Rekognition eye gaze direction detection to monitor driver behavior and identify distractions.
B. Use Amazon SageMaker AI to customize an AI model to monitor driver behavior and identify distractions.
C. Integrate a third-party driver monitoring system with Amazon Rekognition to monitor driver behavior and identify distractions
D. Use Amazon Comprehend to analyze text-based driver feedback and identify distractions.
Show Answer
Correct Answer: A
Explanation: Amazon Rekognition provides built-in facial analysis capabilities, including eye gaze detection, that can be used to detect driver distraction with minimal development and operational overhead. Amazon SageMaker would require building, training, and maintaining a custom model. A third-party integration adds operational complexity. Amazon Comprehend analyzes text, not video or images, so it is not suitable for monitoring driver behavior.
Question 59
A company has significantly increased the amount of data that is stored as .csv files in an Amazon S3 bucket. Data transformation scripts and queries are now taking much longer than they used to take.
An ML engineer must implement a solution to optimize the data for query performance.
Which solution will meet this requirement with the LEAST operational overhead?
A. Configure an AWS Lambda function to split the .csv files into smaller objects in the S3 bucket.
B. Configure an AWS Glue job to drop columns that have string type values and to save the results to the S3 bucket.
C. Configure an AWS Glue extract, transform, and load (ETL) job to convert the .csv files to Apache Parquet format.
D. Configure an Amazon EMR cluster to process the data that is in the S3 bucket.
Show Answer
Correct Answer: C
Explanation: Converting CSV data to Apache Parquet with an AWS Glue ETL job optimizes storage and query performance because Parquet is a compressed, columnar format that reduces I/O and speeds analytics. AWS Glue is a serverless managed ETL service, so it provides the required transformation with minimal operational overhead compared to managing EMR. Splitting CSV files does not address the inefficiency of the file format, and dropping string columns changes the dataset rather than optimizing it.
Question 60
A company is using Amazon SageMaker AI to deploy a new recommendation model for its ecommerce website. The model must use data from all client website interactions as input.
Traffic is variable throughout the day. The company needs to create an inference endpoint for the model.
Which type of inference endpoint will meet these requirements MOST cost-effectively?
A. Batch transform inference endpoint
B. Asynchronous inference endpoint
C. Real-time inference endpoint
D. Serverless inference endpoint
Show Answer
Correct Answer: D
Explanation: The endpoint must serve online recommendations using live interaction data, so batch transform is not appropriate. Asynchronous inference is intended for long-running requests rather than low-latency recommendation serving. Real-time endpoints provide low latency but require provisioned instances. With variable and unpredictable traffic, SageMaker Serverless Inference automatically scales with demand and can scale to zero, making it the most cost-effective choice for this workload.
$19
Get all 235 questions with detailed answers and explanations
Instant download HTML + PDF delivered the moment payment clears.
Secure Stripe checkout we never see or store your card details.
7-day refund if files are defective see our refund policy.
