A company has installed a security camera. The company uses an ML model to evaluate the security camera footage for potential thefts. The company has discovered that the model disproportionately flags people who are members of a specific ethnic group. Which type of bias is affecting the model output?
A. Measurement bias B. Sampling bias C. Observer bias D. Confirmation bias
Answer: B Explanation:Sampling bias is the correct type of bias affecting the model output when itdisproportionately flags people from a specific ethnic group.Sampling Bias:Why Option B is Correct:Why Other Options are Incorrect:
Question # 32
An AI practitioner trained a custom model on Amazon Bedrock by using a training dataset that contains confidential data. The AI practitioner wants to ensure that the custom model does not generate inference responses based on confidential data. How should the AI practitioner prevent responses based on confidential data?
A. Delete the custom model. Remove the confidential data from the training dataset.Retrain the custom model. B. Mask the confidential data in the inference responses by using dynamic data masking. C. Encrypt the confidential data in the inference responses by using Amazon SageMaker. D. Encrypt the confidential data in the custom model by using AWS Key Management Service (AWS KMS).
Answer: A Explanation:When a model is trained on a dataset containing confidential or sensitive data, the modelmay inadvertently learn patterns from this data, which could then be reflected in itsinference responses. To ensure that a model does not generate responses based onconfidential data, the most effective approach is to remove the confidential data from thetraining dataset and then retrain the model. Explanation of Each Option:Option A (Correct):"Delete the custom model. Remove the confidential data fromthe training dataset. Retrain the custom model."This option is correct because itdirectly addresses the core issue: the model has been trained on confidential data.The only way to ensure that the model does not produce inferences based on thisdata is to remove the confidential information from the training dataset and thenretrain the model from scratch. Simply deleting the model and retraining it ensuresthat no confidential data is learned or retained by the model. This approach followsthe best practices recommended by AWS for handling sensitive data when usingmachine learning services like Amazon Bedrock.Option B:"Mask the confidential data in the inference responses by using dynamicdata masking."This option is incorrect because dynamic data masking is typicallyused to mask or obfuscate sensitive data in a database. It does not address thecore problem of the model being trained on confidential data. Masking data ininference responses does not prevent the model from using confidential data itlearned during training.Option C:"Encrypt the confidential data in the inference responses by usingAmazon SageMaker."This option is incorrect because encrypting the inferenceresponses does not prevent the model from generating outputs based onconfidential data. Encryption only secures the data at rest or in transit but does notaffect the model's underlying knowledge or training process.Option D:"Encrypt the confidential data in the custom model by using AWS KeyManagement Service (AWS KMS)."This option is incorrect as well becauseencrypting the data within the model does not prevent the model from generatingresponses based on the confidential data it learned during training. AWS KMS canencrypt data, but it does not modify the learning that the model has alreadyperformed.AWS AI Practitioner References:Data Handling Best Practices in AWS Machine Learning:AWS advisespractitioners to carefully handle training data, especially when it involves sensitiveor confidential information. This includes preprocessing steps like dataanonymization or removal of sensitive data before using it to train machinelearning models.Amazon Bedrock and Model Training Security:Amazon Bedrock providesfoundational models and customization capabilities, but any training involvingsensitive data should follow best practices, such as removing or anonymizingconfidential data to prevent unintended data leakage.
Question # 33
An AI practitioner is using an Amazon Bedrock base model to summarize session chats from the customer service department. The AI practitioner wants to store invocation logs to monitor model input and output data. Which strategy should the AI practitioner use?
A. Configure AWS CloudTrail as the logs destination for the model. B. Enable invocation logging in Amazon Bedrock. C. Configure AWS Audit Manager as the logs destination for the model. D. Configure model invocation logging in Amazon EventBridge.
Answer: B Explanation: Amazon Bedrock provides an option to enable invocation logging to captureand store the input and output data of the models used. This is essential for monitoring andauditing purposes, particularly when handling customer data.Option B (Correct): "Enable invocation logging in Amazon Bedrock":This is thecorrect answer as it directly enables the logging of all model invocations, ensuringtransparency and traceability.Option A:"Configure AWS CloudTrail" is incorrect because CloudTrail logs APIcalls but does not provide specific logging for model inputs and outputs.Option C:"Configure AWS Audit Manager" is incorrect as Audit Manager is usedfor compliance reporting, not specific invocation logging for AI models.Option D:"Configure model invocation logging in Amazon EventBridge" is incorrectas EventBridge is for event-driven architectures, not specifically designed forlogging AI model inputs and outputs.AWS AI Practitioner References:Amazon Bedrock Logging Capabilities:AWS emphasizes using built-in loggingfeatures in Bedrock to maintain data integrity and transparency in modeloperations.
Question # 34
An AI company periodically evaluates its systems and processes with the help of independent software vendors (ISVs). The company needs to receive email message notifications when an ISV's compliance reports become available. Which AWS service can the company use to meet this requirement?
A. AWS Audit Manager B. AWS Artifact C. AWS Trusted Advisor D. AWS Data Exchange
Answer: D Explanation:AWS Data Exchange is a service that allows companies to securely exchange data withthird parties, such as independent software vendors (ISVs). AWS Data Exchange can beconfigured to provide notifications, including email notifications, when new datasets orcompliance reports become available.Option D (Correct): "AWS Data Exchange":This is the correct answer because itenables the company to receive notifications, including email messages, whenISVs' compliance reports are available.Option A:"AWS Audit Manager" is incorrect because it focuses on assessing anorganization's own compliance, not receiving third-party compliance reports.Option B:"AWS Artifact" is incorrect as it provides access to AWS’s compliancereports, not ISVs'.Option C:"AWS Trusted Advisor" is incorrect as it offers optimization and bestpractices guidance, not compliance report notifications.AWS AI Practitioner References:AWS Data Exchange Documentation:AWS explains how Data Exchange allowsorganizations to subscribe to third-party data and receive notifications whenupdates are available.
Question # 35
A company is building a solution to generate images for protective eyewear. The solution must have high accuracy and must minimize the risk of incorrect annotations. Which solution will meet these requirements?
A. Human-in-the-loop validation by using Amazon SageMaker Ground Truth Plus B. Data augmentation by using an Amazon Bedrock knowledge base C. Image recognition by using Amazon Rekognition D. Data summarization by using Amazon QuickSight
Answer: A Explanation: Amazon SageMaker Ground Truth Plus is a managed data labeling servicethat includes human-in-the-loop (HITL) validation. This solution ensures high accuracy byinvolving human reviewers to validate the annotations and reduce the risk of incorrectannotations.Amazon SageMaker Ground Truth Plus:Why Option A is Correct:Why Other Options are Incorrect:Thus,Ais the correct answer for generating high-accuracy images with minimizedannotation risks.
Question # 36
A company has built an image classification model to predict plant diseases from photos of plant leaves. The company wants to evaluate how many images the model classified correctly. Which evaluation metric should the company use to measure the model's performance?
A. R-squared score B. Accuracy C. Root mean squared error (RMSE) D. Learning rate
Answer: B Explanation:Accuracy is the most appropriate metric to measure the performance of an imageclassification model. It indicates the percentage of correctly classified images out of thetotal number of images. In the context of classifying plant diseases from images, accuracywill help the company determine how well the model is performing by showing how manyimages were correctly classified. Option B (Correct): "Accuracy":This is the correct answer because accuracymeasures the proportion of correct predictions made by the model, which issuitable for evaluating the performance of a classification model.Option A:"R-squared score" is incorrect as it is used for regression analysis, notclassification tasks.Option C:"Root mean squared error (RMSE)" is incorrect because it is also usedfor regression tasks to measure prediction errors, not for classification accuracy.Option D:"Learning rate" is incorrect as it is a hyperparameter for training, not aperformance metric.AWS AI Practitioner References:Evaluating Machine Learning Models on AWS:AWS documentation emphasizesthe use of appropriate metrics, like accuracy, for classification tasks.
Question # 37
A company has petabytes of unlabeled customer data to use for an advertisement campaign. The company wants to classify its customers into tiers to advertise and promote the company's products. Which methodology should the company use to meet these requirements?
A. Supervised learning B. Unsupervised learning C. Reinforcement learning D. Reinforcement learning from human feedback (RLHF)
Answer: B Explanation:Unsupervised learning is the correct methodology for classifying customers into tiers whenthe data is unlabeled, as it does not require predefined labels or outputs.Unsupervised Learning:Why Option B is Correct:Why Other Options are Incorrect:
Question # 38
A security company is using Amazon Bedrock to run foundation models (FMs). The company wants to ensure that only authorized users invoke the models. The company needs to identify any unauthorized access attempts to set appropriate AWS Identity and Access Management (IAM) policies and roles for future iterations of the FMs. Which AWS service should the company use to identify unauthorized users that are trying to access Amazon Bedrock?
A. AWS Audit Manager B. AWS CloudTrail C. Amazon Fraud Detector D. AWS Trusted Advisor
Answer: B Explanation:AWS CloudTrail is a service that enables governance, compliance, and operational andrisk auditing of your AWS account. It tracks API calls and identifies unauthorized accessattempts to AWS resources, including Amazon Bedrock.AWS CloudTrail:Why Option B is Correct:Why Other Options are Incorrect:Thus,Bis the correct answer for identifying unauthorized users attempting to accessAmazon Bedrock.
Question # 39
A company is using an Amazon Bedrock base model to summarize documents for an internal use case. The company trained a custom model to improve the summarization quality. Which action must the company take to use the custom model through Amazon Bedrock?
A. Purchase Provisioned Throughput for the custom model. B. Deploy the custom model in an Amazon SageMaker endpoint for real-time inference. C. Register the model with the Amazon SageMaker Model Registry. D. Grant access to the custom model in Amazon Bedrock.
Answer: B Explanation:To use a custom model that has been trained to improve summarization quality, thecompany must deploy the model on an Amazon SageMaker endpoint. This allows themodel to be used for real-time inference through Amazon Bedrock or other AWS services.By deploying the model in SageMaker, the custom model can be accessedprogrammatically via API calls, enabling integration with Amazon Bedrock.Option B (Correct): "Deploy the custom model in an Amazon SageMaker endpointfor real-time inference":This is the correct answer because deploying the model onSageMaker enables it to serve real-time predictions and be integrated with Amazon Bedrock.Option A:"Purchase Provisioned Throughput for the custom model" is incorrectbecause provisioned throughput is related to database or storage services, notmodel deployment.Option C:"Register the model with the Amazon SageMaker Model Registry" isincorrect because while the model registry helps with model management, it doesnot make the model accessible for real-time inference.Option D:"Grant access to the custom model in Amazon Bedrock" is incorrectbecause Bedrock does not directly manage custom model access; it relies ondeployed endpoints like those in SageMaker.AWS AI Practitioner References:Amazon SageMaker Endpoints:AWS recommends deploying models toSageMaker endpoints to use them for real-time inference in various applications.
Question # 40
An accounting firm wants to implement a large language model (LLM) to automate document processing. The firm must proceed responsibly to avoid potential harms.What should the firm do when developing and deploying the LLM? (Select TWO.)
A. Include fairness metrics for model evaluation. B. Adjust the temperature parameter of the model. C. Modify the training data to mitigate bias. D. Avoid overfitting on the training data. E. Apply prompt engineering techniques.
Answer: A,C Explanation:To implement a large language model (LLM) responsibly, the firm should focus on fairnessand mitigating bias, which are critical for ethical AI deployment.A. Include Fairness Metrics for Model Evaluation:C. Modify the Training Data to Mitigate Bias:Why Other Options are Incorrect: