Amazon AIF-C01 Online Practice Questions and Exam Preparation

Amazon AIF-C01 Online Questions & Answers

• Question 241:

  HOTSPOT

  A company has developed a large language model (LLM) and wants to make the LLM available to multiple internal teams. The company needs to select the appropriate inference mode for each team.

  Select the correct inference mode from the following list for each use case. Each inference mode should be selected one or more times.

  

  

  Explanation:

  The company's chatbot needs predictions from the LLM to understand users' intent with minimal latency - Real-time inference

  This scenario requires Real-time inference, as chatbots typically need immediate responses with low latency to provide an interactive user experience.

  A data processing job needs to query the LLM to process gigabytes of text files on weekends - Batch transform

  This use case is best suited for Batch transform. Since the data processing job involves handling large amounts of data, and it is scheduled for weekends, batch processing can handle these large volumes efficiently.

  The company's engineering team needs to create an API that can process small pieces of text content and provide low-latency predictions - Real-time inference Real-time inference is the correct choice here, as the API needs to process text quickly and provide immediate responses, making real-time inference appropriate.

• Question 242:

  A healthcare company is building an AI solution to predict patient readmission within 30 days of patient discharge. The company has trained a model on historical patient data including medical history, demographics, and treatment specifications, to provide readmission predictions in real time. Which task describes AI model inference in this scenario?

  A. Gather historical patient readmission data.
  B. Use appropriate metrics and assess model performance.
  C. Use data to identify patient patterns and correlations.
  D. Use a trained model to predict patient readmission.
  D. Use a trained model to predict patient readmission.

  ---

  Model inference refers to the process of applying a trained model to new data--in this case, using the trained model to make real-time predictions about whether a patient will be readmitted within 30 days of discharge.

• Question 243:

  An online learning company with large volumes of education materials wants to use enterprise search. Which AWS service meets these requirements?

  A. Amazon Comprehend
  B. Amazon Textract
  C. Amazon Kendra
  D. Amazon Personalize
  C. Amazon Kendra

  ---

  Amazon Kendra is an intelligent enterprise search service that enables organizations to search large volumes of unstructured data, such as educational materials, making it ideal for enterprise search needs.

• Question 244:

  A company wants to make a chatbot to help customers. The chatbot will help solve technical problems without human intervention. The company chose a foundation model (FM) for the chatbot. The chatbot needs to produce responses that adhere to company tone. Which solution meets these requirements?

  A. Set a low limit on the number of tokens the FM can produce.
  B. Use batch inferencing to process detailed responses.
  C. Experiment and refine the prompt until the FM produces the desired responses.
  D. Define a higher number for the temperature parameter.
  C. Experiment and refine the prompt until the FM produces the desired responses.

  ---

  Experimenting and refining the prompt is the best approach to ensure that the chatbot using a foundation model (FM) produces responses that adhere to the company's tone.

• Question 245:

  A bank has ne-tuned a large language model (LLM) to expedite the loan approval process. During an external audit of the model, the company discovered that the model was approving loans at a faster pace for a specific demographic than for other demographics. How should the bank fix this issue MOST cost-effectively?

  A. Include more diverse training data. Fine-tune the model again by using the new data.
  B. Use Retrieval Augmented Generation (RAG) with the ne-tuned model.
  C. Use AWS Trusted Advisor checks to eliminate bias.
  D. Pre-train a new LLM with more diverse training data.
  A. Include more diverse training data. Fine-tune the model again by using the new data.

  ---

  The issue described is a case of bias in the model's decision-making process, where the model is showing a preference for a specific demographic. The most cost-effective way to address this is to include more diverse training data that better represents all demographics. By fine-tuning the model with this more diverse data, you can help ensure that the model treats all demographic groups fairly and does not exhibit biased behavior. Fine-tuning the model with updated data ensures that the model learns from a more representative sample, improving fairness in its predictions without needing to completely retrain the model from scratch.

• Question 246:

  A company wants to train a model but must ensure sensitive training data never leaves its AWS Region. Which approach meets these requirements?

  A. Use Amazon CloudFront with encryption
  B. Use Bedrock API with Regional endpoints and PrivateLink
  C. Export data to a local workstation for preprocessing
  D. Use Amazon Macie to mask sensitive fields before export
  B. Use Bedrock API with Regional endpoints and PrivateLink

  ---

  Amazon Bedrock operates regionally and, combined with PrivateLink, ensures all data stays within the customer's VPC and Region. This satisfies data residency and compliance requirements.

• Question 247:

  A company needs to choose a model from Amazon Bedrock to use internally. The company must identify a model that generates responses in a style that the company's employees prefer. What should the company do to meet these requirements?

  A. Evaluate the models by using built-in prompt datasets.
  B. Evaluate the models by using a human workforce and custom prompt datasets.
  C. Use public model leaderboards to identify the model.
  D. Use the model InvocationLatency runtime metrics in Amazon CloudWatch when trying models.
  B. Evaluate the models by using a human workforce and custom prompt datasets.

  ---

  To determine which model generates responses in a style that the company's employees prefer, the best approach is to use a human workforce to evaluate the models with custom prompt datasets. This method allows for subjective evaluation based on the specific stylistic preferences of the company's employees, which cannot be effectively assessed through automated methods or pre-built datasets.

  Option B (Correct): "Evaluate the models by using a human workforce and custom prompt datasets":This is the correct answer as it directly involves human judgment to evaluate the style and quality of the responses, aligning with employee preferences.

  Option A:"Evaluate the models by using built-in prompt datasets" is incorrect because built-in datasets may not capture the company's specific stylistic requirements.

  Option C:"Use public model leaderboards to identify the model" is incorrect as leaderboards typically measure model performance on standard benchmarks, not on stylistic preferences.

  Option D:"Use the model InvocationLatency runtime metrics in Amazon CloudWatch" is incorrect because latency metrics do not provide any information about the style of the model's responses.

  References:

  Model Evaluation Techniques on AWS:AWS suggests using human evaluators to assess qualitative aspects of model outputs, such as style and tone, to ensure alignment with organizational preferences

• Question 248:

  A media company wants to analyze viewer behavior and demographics to recommend personalized content. The company wants to deploy a customized ML model in its production environment. The company also wants to observe if the model quality drifts over time. Which AWS service or feature meets these requirements?

  A. Amazon Rekognition
  B. Amazon SageMaker Clarify
  C. Amazon Comprehend
  D. Amazon SageMaker Model Monitor
  D. Amazon SageMaker Model Monitor

  ---

  Amazon SageMaker Model Monitor continuously tracks deployed ML models in production to detect data drift, model drift, and quality degradation over time. This is essential for ensuring that the recommendation model remains accurate as viewer behavior and demographics change. Model Monitor helps detect anomalies and provides alerts when model performance deviates from expected trends, allowing the company to take corrective action.

• Question 249:

  A company has an ML model. The company wants to know how the model makes predictions. Which term refers to understanding model predictions?

  A. Model interpretability
  B. Model training
  C. Model interoperability
  D. Model performance
  A. Model interpretability

  ---

  Model interpretability refers to understanding how a model makes its predictions, providing insights into the reasoning behind the model's outputs.

• Question 250:

  A company has developed a generative AI application by using Amazon Bedrock. The application produces catalogs that are updated weekly. The company needs to select a pricing model for weekly catalog updates.

  Which Amazon Bedrock pricing model meets these requirements MOST cost-effectively?

  A. On-Demand
  B. Provisioned Throughput
  C. Batch
  D. Model evaluation
  C. Batch

  ---

  Explanation

  Batch is the most cost-effective choice for weekly catalog updates because it is designed for asynchronous, non-real-time workloads, and AWS states that Amazon Bedrock batch inference can process large jobs offline at lower cost, including pricing at up to 50% of on-demand inference for supported models.