Amazon AIF-C01 Online Practice Questions and Exam Preparation

Amazon AIF-C01 Online Questions & Answers

• Question 101:

  A financial company wants to flag all credit card activity as possibly fraudulent or non-fraudulent based on transaction data. Which type of ML model meets these requirements?

  A. Regression
  B. Diffusion
  C. Binary classification
  D. Multi-class classification
  C. Binary classification

  ---

  Binary classification models are designed to distinguish between two classes - fraudulent versus non- fraudulent transactions - making them the appropriate choice for this use case.

• Question 102:

  Which option is a use case for generative AI models?

  A. Improving network security by using intrusion detection systems
  B. Creating photorealistic images from text descriptions for digital marketing
  C. Enhancing database performance by using optimized indexing
  D. Analyzing financial data to forecast stock market trends
  B. Creating photorealistic images from text descriptions for digital marketing

  ---

  Generative AI models are used to create new content based on existing data. One common use case is generating photorealistic images from text descriptions, which is particularly useful in digital marketing, where visual content is key to engaging potential customers.

  Option B (Correct): "Creating photorealistic images from text descriptions for digital marketing":This is the correct answer because generative AI models, like those offered by Amazon Bedrock, can create images based on text descriptions, making them highly valuable for generating marketing materials.

  Option A:"Improving network security by using intrusion detection systems" is incorrect because this is a use case for traditional machine learning models, not generative AI.

  Option C:"Enhancing database performance by using optimized indexing" is incorrect as it is unrelated to generative AI.

  Option D:"Analyzing financial data to forecast stock market trends" is incorrect because it typically involves predictive modeling rather than generative AI.

  References:

  Use Cases for Generative AI Models on AWS:AWS highlights the use of generative AI for creative content generation, including image creation, text generation, and more, which is suited for digital marketing applications.

• Question 103:

  A company's large language model (LLM) is experiencing hallucinations. How can the company decrease hallucinations?

  A. Set up Agents for Amazon Bedrock to supervise the model training.
  B. Use data pre-processing and remove any data that causes hallucinations.
  C. Decrease the temperature inference parameter for the model.
  D. Use a foundation model (FM) that is trained to not hallucinate.
  C. Decrease the temperature inference parameter for the model.

  ---

  In the context of large language models (LLMs), the temperature inference parameter controls the randomness of the model's output. Lowering the temperature reduces randomness and makes the model's responses more deterministic and focused on the most likely predictions. By decreasing the temperature, the likelihood of hallucinations (when the model generates incorrect or nonsensical information) is reduced, as the model relies more on high-probability outputs rather than exploring less likely possibilities.

• Question 104:

  A company deployed a Retrieval Augmented Generation (RAG) application on Amazon Bedrock that gathers financial news to distribute in daily newsletters. Users have recently reported politically influenced ideas in the newsletters. Which Amazon Bedrock guardrail can identify and filter this content?

  A. Word filters
  B. Denied topics
  C. Sensitive information filters
  D. Content filters
  B. Denied topics

  ---

  The Denied topics guardrail lets you define high-level categories (such as politics) that the model must avoid. Bedrock enforces these rules during generation, filtering out any content related to politically influenced ideas.

• Question 105:

  A company wants to classify images of different objects based on custom features extracted from a dataset. Which solution will meet this requirement with the LEAST development effort?

  A. Use traditional ML algorithms with custom features extracted from the dataset.
  B. Use a pre-trained deep learning model. Fine-tune the model on the dataset.
  C. Use a generative adversarial network (GAN) model to classify the images.
  D. Use a support vector machine (SVM) with manually engineered features for classification.
  B. Use a pre-trained deep learning model. Fine-tune the model on the dataset.

  ---

  Fine-tuning a pre-trained deep learning model leverages transfer learning, reducing development effort because the model already captures general image features. Only minimal adjustments with the custom dataset are needed to achieve accurate classification.

• Question 106:

  HOTSPOT

  A company wants to improve multiple ML models.

  Select the correct technique from the following list of use cases. Each technique should be selected one time or not at all.

  

  

  Explanation:

  Enhancing the capabilities of a large language model (LLM) by using external sources -> Retrieval Augmented Generation (RAG)

  Querying a model to generalize and make predictions on unseen tasks -> Zero-shot learning Querying a model with a limited amount of data for new tasks -> Few-shot learning

  RAG improves LLMs by incorporating information from external data sources at query time.

  Zero-shot learning allows a model to perform tasks it wasn't explicitly trained on, using generalization.

  Few-shot learning enables a model to handle new tasks with only a small number of examples.

• Question 107:

  A financial institution wants to prevent its AI model from generating investment advice that violates compliance rules. The model is hosted in Amazon Bedrock. The company needs an automated mechanism to block outputs containing restricted content. Which solution should the company use?

  A. Amazon CloudWatch Logs
  B. Amazon Bedrock Guardrails
  C. Amazon SageMaker Clarify
  D. AWS Artifact
  B. Amazon Bedrock Guardrails

  ---

  Amazon Bedrock Guardrails allow organizations to define prohibited content categories, blocked keywords, and safety rules for both user inputs and model outputs. By enforcing compliance constraints at runtime, guardrails prevent the model from generating restricted financial guidance.

• Question 108:

  A company wants to identify sensitive data such as personally identifiable information (PII) that is stored in Amazon S3 before using the data in an AI workload.

  Which AWS service should the company use?

  A. Amazon Inspector
  B. Amazon Macie
  C. Amazon Polly
  D. AWS Budgets
  B. Amazon Macie

  ---

  Explanation

  Amazon Macie is the correct answer because it helps discover and protect sensitive data in Amazon S3, including types of personally identifiable information.

  Option B (Correct): "Amazon Macie": This is correct because it is designed for sensitive data discovery in S3.

  Option A: "Amazon Inspector" is incorrect because it focuses on vulnerability management.

  Option C: "Amazon Polly" is incorrect because it is a text-to-speech service.

  Option D: "AWS Budgets" is incorrect because it is used for financial monitoring.

• Question 109:

  A company is exploring Amazon Nova models in Amazon Bedrock. The company needs a multimodal model that supports multiple languages. Which Nova model will meet these requirements MOST cost-effectively?

  A. Nova Lite
  B. Nova Pro
  C. Nova Canvas
  D. Nova Reel
  A. Nova Lite

  ---

  Nova Lite is a multimodal model in Amazon Bedrock that supports multiple languages and is designed to be the most cost-effective option among the Nova models, making it suitable for organizations seeking efficient, scalable, and economical AI solutions.

• Question 110:

  Which metric measures the runtime efficiency of operating AI models?

  A. Customer satisfaction score (CSAT)
  B. Training time for each epoch
  C. Average response time
  D. Number of training instances
  C. Average response time

  ---

  The average response time is the correct metric for measuring the runtime efficiency of operating AI models.