Amazon AIF-C01 Online Practice Questions and Exam Preparation

Amazon AIF-C01 Online Questions & Answers

• Question 151:

  HOTSPOT

  A company wants to develop ML applications to improve business operations and e ciency.

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

  

  

  Explanation:

  Binary classification: Supervised learning

  Binary classification involves predicting one of two possible outcomes, requiring labeled training data.

  Multi-class classification: Supervised learning

  Multi-class classification extends binary classification to predict one of several classes, also requiring labeled data.

  K-means clustering: Unsupervised learning

  K-means clustering is used to group data points into clusters without requiring labeled data.

  Dimensionality reduction: Unsupervised learning

  Dimensionality reduction techniques, like PCA, reduce the number of features in the dataset without needing labeled data.

• Question 152:

  A company wants to compare two foundation models for a summarization use case. The company wants people to judge which model produces summaries that are more useful for employees.

  Which evaluation approach should the company use?

  A. Human evaluation with custom prompts
  B. Only measure invocation latency
  C. Only compare the number of model parameters
  D. Only increase the context window
  A. Human evaluation with custom prompts

  ---

  Explanation

  Human evaluation with custom prompts is the correct answer because usefulness is a qualitative requirement that depends on the preferences and needs of the company's users. Human reviewers are best suited to assess that kind of output quality.

  Option A (Correct): "Human evaluation with custom prompts": This is correct because people can directly assess usefulness and quality.

  Option B: "Only measure invocation latency" is incorrect because speed does not measure summary usefulness.

  Option C: "Only compare the number of model parameters" is incorrect because size alone does not determine output quality.

  Option D: "Only increase the context window" is incorrect because that is not an evaluation method.

• Question 153:

  A news agency publishes articles in English. The agency wants to make articles available in other languages. Which solution meets these requirements?

  A. Add Amazon Transcribe to the company's website.
  B. Use the Amazon Translate real-time translation feature.
  C. Add Amazon Personalize to the company's website.
  D. Use the Amazon Textract real-time document processing feature.
  B. Use the Amazon Translate real-time translation feature.

  ---

  Amazon Translate provides real-time translation of text, making it suitable for automatically translating articles from English into other languages for publication.

• Question 154:

  A financial institution is building an AI solution to make loan approval decisions by using a foundation model (FM). For security and audit purposes, the company needs the AI solution's decisions to be explainable. Which factor relates to the explainability of the AI solution's decisions?

  A. Model complexity
  B. Training time
  C. Number of hyperparameters
  D. Deployment time
  A. Model complexity

  ---

  Model complexity plays a significant role in the explainability of an AI solution. Simpler models tend to be more explainable because it is easier to understand how they make decisions. For example, linear regression or decision trees are generally easier to interpret compared to more complex models like deep neural networks.

  In contrast, highly complex models, such as deep learning models, may provide very accurate results but are often considered "black boxes," meaning their decision-making process is not easily interpretable. This lack of transparency makes it difficult to explain their decisions in a way that satisfies regulatory requirements or customer understanding.

• Question 155:

  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).
  A. Delete the custom model. Remove the confidential data from the training dataset. Retrain the custom model.

  ---

  When a model is trained on a dataset containing confidential or sensitive data, the model may inadvertently learn patterns from this data, which could then be reflected in its inference responses. To ensure that a model does not generate responses based on confidential data, the most effective approach is to remove the confidential data from the training dataset and then retrain the model.

  of Each Option:

  Option A (Correct):"Delete the custom model. Remove the confidential data from the training dataset. Retrain the custom model."This option is correct because it directly 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 this data is to remove the confidential information from the training dataset and then retrain the model from scratch. Simply deleting the model and retraining it ensures that no confidential

  data is learned or retained by the model. This approach follows the best practices recommended by AWS for handling sensitive data when using machine learning services like Amazon Bedrock.

  Option B:"Mask the confidential data in the inference responses by using dynamic data masking."This option is incorrect because dynamic data masking is typically used to mask or obfuscate sensitive data in a database. It does not address the core problem of the model being trained on confidential data. Masking data in inference responses does not prevent the model from using confidential data it learned during training.

  Option C:"Encrypt the confidential data in the inference responses by using Amazon SageMaker."This option is incorrect because encrypting the inference responses does not prevent the model from generating outputs based on confidential data. Encryption only secures the data at rest or in transit but does not affect the model's underlying knowledge or training process.

  Option D:"Encrypt the confidential data in the custom model by using AWS Key Management Service (AWS KMS)."This option is incorrect as well because encrypting the data within the model does not prevent the model from generating responses based on the confidential data it learned during training. AWS KMS can encrypt data, but it does not modify the learning that the model has already performed.

  References:

  Data Handling Best Practices in AWS Machine Learning:AWS advises practitioners to carefully handle training data, especially when it involves sensitive or confidential information. This includes preprocessing steps like data anonymization or removal of sensitive data before using it to train machine learning models.

  Amazon Bedrock and Model Training Security:Amazon Bedrock provides foundational models and customization capabilities, but any training involving sensitive data should follow best practices, such as removing or anonymizing confidential data to prevent unintended data leakage.

• Question 156:

  A company is building a new generative AI chatbot. The chatbot uses an Amazon Bedrock foundation model (FM) to generate responses. During testing, the company notices that the chatbot is prone to prompt injection attacks. What can the company do to secure the chatbot with the LEAST implementation effort?

  A. Fine-tune the FM to avoid harmful responses.
  B. Use Amazon Bedrock Guardrails content filters and denied topics.
  C. Change the FM to a more secure FM.
  D. Use chain-of-thought prompting to produce secure responses.
  B. Use Amazon Bedrock Guardrails content filters and denied topics.

  ---

  Using Amazon Bedrock Guardrails' content filters and denied topics is the quickest and least effort solution to mitigate prompt injection attacks. These guardrails can be configured without model retraining, helping to automatically filter or block unsafe prompts and responses.

• Question 157:

  Which metric is used to evaluate the performance of foundation models (FMs) for text summarization tasks?

  A. F1 score
  B. Bilingual Evaluation Understudy (BLEU) score
  C. Accuracy
  D. Mean squared error (MSE)
  A. F1 score

  ---

  Text summarization quality is most commonly assessed using ROUGE metrics, which report recall, precision, and an F1-score that balances the two. The ROUGE-F1 (often simply called the F1 score in this context) measures how well the model's summary overlaps with reference summaries, making it the standard choice for evaluating foundation models on summarization tasks.

• Question 158:

  A company wants an application to generate several different slogan ideas from the same prompt. The company wants the model to produce more varied outputs.

  Which inference parameter adjustment should the company make?

  A. Decrease the temperature
  B. Increase the temperature
  C. Decrease the number of input tokens
  D. Use batch inference
  B. Increase the temperature

  ---

  Explanation

  Increasing the temperature is the correct answer because higher temperature values generally make model outputs more diverse and less deterministic. This is useful when a business wants multiple creative variations.

  Option B (Correct): "Increase the temperature": This is correct because a higher temperature encourages more varied generations.

  Option A: "Decrease the temperature" is incorrect because it usually makes outputs more predictable and repetitive.

  Option C: "Decrease the number of input tokens" is incorrect because that does not directly control creativity.

  Option D: "Use batch inference" is incorrect because batch inference relates to processing mode, not output diversity.

• Question 159:

  An AI practitioner wants to evaluate ML models. The AI practitioner wants to provide explanations of model predictions to customers and stakeholders. Which AWS service or feature will meet these requirements?

  A. Amazon QuickSight
  B. Amazon Comprehend
  C. AWS Trusted Advisor
  D. Amazon SageMaker Clarify
  D. Amazon SageMaker Clarify

  ---

  Amazon SageMaker Clarify provides tools to explain model predictions, allowing practitioners to generate and share interpretability reports with customers and stakeholders, improving transparency and trust in ML models.

• Question 160:

  A company built a deep learning model for object detection and deployed the model to production.

  Which AI process occurs when the model analyzes a new image to identify objects?

  A. Training
  B. Inference
  C. Model deployment
  D. Bias correction
  B. Inference

  ---

  Inference is the correct answer because it is the AI process that occurs when a deployed model analyzes new data (such as an image) to make predictions or identify objects.