A machine learning (ML) specialist wants to bring a custom training algorithm to Amazon SageMaker. The ML specialist implements the algorithm in a Docker container that is supported by SageMaker.
How should the ML specialist package the Docker container so that SageMaker can launch the training correctly?

A. Specify the server argument in the ENTRYPOINT instruction in the Dockerfile.
B. Specify the training program in the ENTRYPOINT instruction in the Dockerfile.
C. Include the path to the training data in the docker build command when packaging the container.
D. Use a COPY instruction in the Dockerfile to copy the training program to the /opt/ml/train directory.

A data engineer needs to provide a team of data scientists with the appropriate dataset to run machine learning training jobs. The data will be stored in Amazon S3. The data engineer is obtaining the data from an Amazon Redshift database and is using join queries to extract a single tabular dataset. A portion of the schema is as follows:
TransactionTimestamp (Timestamp)
CardName (Varchar)
CardNo (Varchar)
The data engineer must provide the data so that any row with a CardNo value of NULL is removed. Also, the TransactionTimestamp column must be separated into a TransactionDate column and a TransactionTime column. Finally, the CardName column must be renamed to NameOnCard.
The data will be extracted on a monthly basis and will be loaded into an S3 bucket. The solution must minimize the effort that is needed to set up infrastructure for the ingestion and transformation. The solution also must be automated and must minimize the load on the Amazon Redshift cluster.
Which solution meets these requirements?

A. Set up an Amazon EMR cluster. Create an Apache Spark job to read the data from the Amazon Redshift cluster and transform the data. Load the data into the S3 bucket. Schedule the job to run monthly.
B. Set up an Amazon EC2 instance with a SQL client tool, such as SQL Workbench/J, to query the data from the Amazon Redshift cluster directly Export the resulting dataset into a file. Upload the file into the S3 bucket. Perform these tasks monthly.
C. Set up an AWS Glue job that has the Amazon Redshift cluster as the source and the S3 bucket as the destination. Use the built-in transforms Filter, Map, and RenameField to perform the required transformations. Schedule the job to run monthly.
D. Use Amazon Redshift Spectrum to run a query that writes the data directly to the S3 bucket. Create an AWS Lambda function to run the query monthly.

A data engineer is using AWS Glue to create optimized, secure datasets in Amazon S3. The data science team wants the ability to access the ETL scripts directly from Amazon SageMaker notebooks within a VPC. After this setup is complete, the data science team wants the ability to run the AWS Glue job and invoke the SageMaker training job.
Which combination of steps should the data engineer take to meet these requirements? (Choose three.)

A. Create a SageMaker development endpoint in the data science team's VPC.
B. Create an AWS Glue development endpoint in the data science team's VPC.
C. Create SageMaker notebooks by using the AWS Glue development endpoint.
D. Create SageMaker notebooks by using the SageMaker console.
E. Attach a decryption policy to the SageMaker notebooks.
F. Create an IAM policy and an IAM role for the SageMaker notebooks.

A retail company collects customer comments about its products from social media, the company website, and customer call logs. A team of data scientists and engineers wants to find common topics and determine which products the customers are referring to in their comments. The team is using natural language processing (NLP) to build a model to help with this classification.
Each product can be classified into multiple categories that the company defines. These categories are related but are not mutually exclusive. For example, if there is mention of "Sample Yogurt" in the document of customer comments, then "Sample Yogurt" should be classified as "yogurt," "snack," and "dairy product."
The team is using Amazon Comprehend to train the model and must complete the project as soon as possible.
Which functionality of Amazon Comprehend should the team use to meet these requirements?

A. Custom classification with multi-class mode
B. Custom classification with multi-label mode
C. Custom entity recognition
D. Built-in models

An energy company has wind turbines, weather stations, and solar panels that generate telemetry data. The company wants to perform predictive maintenance on these devices. The devices are in various locations and have unstable internet connectivity.
A team of data scientists is using the telemetry data to perform machine learning (ML) to conduct anomaly detection and predict maintenance before the devices start to deteriorate. The team needs a scalable, secure, high-velocity data ingestion mechanism. The team has decided to use Amazon S3 as the data storage location.
Which approach meets these requirements?

A. Ingest the data by using an HTTP API call to a web server that is hosted on Amazon EC2. Set up EC2 instances in an Auto Scaling configuration behind an Elastic Load Balancer to load the data into Amazon S3.
B. Ingest the data over Message Queuing Telemetry Transport (MQTT) to AWS IoT Core. Set up a rule in AWS IoT Core to use Amazon Kinesis Data Firehose to send data to an Amazon Kinesis data stream that is configured to write to an S3 bucket.
C. Ingest the data over Message Queuing Telemetry Transport (MQTT) to AWS IoT Core. Set up a rule in AWS IoT Core to direct all MQTT data to an Amazon Kinesis Data Firehose delivery stream that is configured to write to an S3 bucket.
D. Ingest the data over Message Queuing Telemetry Transport (MQTT) to Amazon Kinesis data stream that is configured to write to an S3 bucket.

A data scientist is evaluating a GluonTS on Amazon SageMaker DeepAR model. The evaluation metrics on the test set indicate that the coverage score is 0.489 and 0.889 at the 0.5 and 0.9 quantiles, respectively.
What can the data scientist reasonably conclude about the distributional forecast related to the test set?

A. The coverage scores indicate that the distributional forecast is poorly calibrated. These scores should be approximately equal to each other at all quantiles.
B. The coverage scores indicate that the distributional forecast is poorly calibrated. These scores should peak at the median and be lower at the tails.
C. The coverage scores indicate that the distributional forecast is correctly calibrated. These scores should always fall below the quantile itself.
D. The coverage scores indicate that the distributional forecast is correctly calibrated. These scores should be approximately equal to the quantile itself.

A data scientist is reviewing customer comments about a company's products. The data scientist needs to present an initial exploratory analysis by using charts and a word cloud. The data scientist must use feature engineering techniques to prepare this analysis before starting a natural language processing (NLP) model.
Which combination of feature engineering techniques should the data scientist use to meet these requirements? (Choose two.)

A. Named entity recognition
B. Coreference
C. Stemming
D. Term frequency-inverse document frequency (TF-IDF)
E. Sentiment analysis

A real-estate company is launching a new product that predicts the prices of new houses. The historical data for the properties and prices is stored in .csv format in an Amazon S3 bucket. The data has a header, some categorical fields, and some missing values. The company's data scientists have used Python with a common open-source library to fill the missing values with zeros. The data scientists have dropped all of the categorical fields and have trained a model by using the open-source linear regression algorithm with the default parameters.
The accuracy of the predictions with the current model is below 50%. The company wants to improve the model performance and launch the new product as soon as possible.
Which solution will meet these requirements with the LEAST operational overhead?

A. Create a service-linked role for Amazon Elastic Container Service (Amazon ECS) with access to the S3 bucket. Create an ECS cluster that is based on an AWS Deep Learning Containers image. Write the code to perform the feature engineering. Train a logistic regression model for predicting the price, pointing to the bucket with the dataset. Wait for the training job to complete. Perform the inferences.
B. Create an Amazon SageMaker notebook with a new IAM role that is associated with the notebook. Pull the dataset from the S3 bucket. Explore different combinations of feature engineering transformations, regression algorithms, and hyperparameters. Compare all the results in the notebook, and deploy the most accurate configuration in an endpoint for predictions.
C. Create an IAM role with access to Amazon S3, Amazon SageMaker, and AWS Lambda. Create a training job with the SageMaker built-in XGBoost model pointing to the bucket with the dataset. Specify the price as the target feature. Wait for the job to complete. Load the model artifact to a Lambda function for inference on prices of new houses.
D. Create an IAM role for Amazon SageMaker with access to the S3 bucket. Create a SageMaker AutoML job with SageMaker Autopilot pointing to the bucket with the dataset. Specify the price as the target attribute. Wait for the job to complete. Deploy the best model for predictions.

A company has an ecommerce website with a product recommendation engine built in TensorFlow. The recommendation engine endpoint is hosted by Amazon SageMaker. Three compute-optimized instances support the expected peak load of the website.
Response times on the product recommendation page are increasing at the beginning of each month. Some users are encountering errors. The website receives the majority of its traffic between 8 AM and 6 PM on weekdays in a single time zone.
Which of the following options are the MOST effective in solving the issue while keeping costs to a minimum? (Choose two.)

A. Configure the endpoint to use Amazon Elastic Inference (EI) accelerators.
B. Create a new endpoint configuration with two production variants.
C. Configure the endpoint to automatically scale with the InvocationsPerInstance metric.
D. Deploy a second instance pool to support a blue/green deployment of models.
E. Reconfigure the endpoint to use burstable instances.

A manufacturing company wants to use machine learning (ML) to automate quality control in its facilities. The facilities are in remote locations and have limited internet connectivity. The company has 20 of training data that consists of labeled images of defective product parts. The training data is in the corporate on-premises data center.
The company will use this data to train a model for real-time defect detection in new parts as the parts move on a conveyor belt in the facilities. The company needs a solution that minimizes costs for compute infrastructure and that maximizes the scalability of resources for training. The solution also must facilitate the company's use of an ML model in the low-connectivity environments.
Which solution will meet these requirements?

A. Move the training data to an Amazon S3 bucket. Train and evaluate the model by using Amazon SageMaker. Optimize the model by using SageMaker Neo. Deploy the model on a SageMaker hosting services endpoint.
B. Train and evaluate the model on premises. Upload the model to an Amazon S3 bucket. Deploy the model on an Amazon SageMaker hosting services endpoint.
C. Move the training data to an Amazon S3 bucket. Train and evaluate the model by using Amazon SageMaker. Optimize the model by using SageMaker Neo. Set up an edge device in the manufacturing facilities with AWS IoT Greengrass. Deploy the model on the edge device.
D. Train the model on premises. Upload the model to an Amazon S3 bucket. Set up an edge device in the manufacturing facilities with AWS IoT Greengrass. Deploy the model on the edge device.