AIF-C01 practice questions 121–130

1. Question 121 · 417 in bank

   A company is building an ML model. The company collected new data and analyzed the data by creating a correlation matrix, calculating statistics, and visualizing the data. Which stage of the ML pipeline is the company currently in?

   • AData pre-processing
   • BFeature engineering
   • CExploratory data analysis
   • DHyperparameter tuning

   Show answer

   Answer: C

   Explanation

   C: Exploratory data analysis Explanation: Exploratory Data Analysis (EDA) involves examining and summarizing data to understand its underlying structure, detect patterns, identify relationships (e.g., via a correlation matrix), and highlight any anomalies. The company's activities, such as creating a correlation matrix, calculating statistics, and visualizing the data, are typical tasks performed during EDA. Why not the other options? A: Data pre-processing: Data pre-processing involves cleaning and preparing data for modeling, such as handling missing values, scaling features, or encoding categorical data. While pre-processing may follow EDA, the tasks described in the question focus on analysis rather than preparation.

   Source: Adapted from community discussion — Moon

   Next question

2. Question 122 · 417 in bank

   A company has documents that are missing some words because of a database error. The company wants to build an ML model that can suggest potential words to fill in the missing text. Which type of model meets this requirement?

   • ATopic modeling
   • BClustering models
   • CPrescriptive ML models
   • DBERT-based models

   Show answer

   Answer: D

   Explanation

   BERT (Bidirectional Encoder Representations from Transformers) is a language model designed to understand context in text by considering both the left and right sides of a word. BERT-based models are well-suited for filling in missing words in sentences due to their ability to predict masked words in a given text. This makes them ideal for tasks that require filling in missing information within text data.

   Source: Adapted from community discussion — dehkon

   Next question

3. Question 123 · 417 in bank

   A company wants to display the total sales for its top-selling products across various retail locations in the past 12 months. Which AWS solution should the company use to automate the generation of graphs?

   • AAmazon Q in Amazon EC2
   • BAmazon Q Developer
   • CAmazon Q in Amazon QuickSight
   • DAmazon Q in AWS Chatbot

   Show answer

   Answer: C

   Next question

4. Question 124 · 417 in bank

   A company is building a chatbot to improve user experience. The company is using a large language model (LLM) from Amazon Bedrock for intent detection. The company wants to use few-shot learning to improve intent detection accuracy. Which additional data does the company need to meet these requirements?

   • APairs of chatbot responses and correct user intents
   • BPairs of user messages and correct chatbot responses
   • CPairs of user messages and correct user intents
   • DPairs of user intents and correct chatbot responses

   Show answer

   Answer: C

   Next question

5. Question 125 · 417 in bank

   A company is using few-shot prompting on a base model that is hosted on Amazon Bedrock. The model currently uses 10 examples in the prompt. The model is invoked once daily and is performing well. The company wants to lower the monthly cost. Which solution will meet these requirements?

   • ACustomize the model by using fine-tuning.
   • BDecrease the number of tokens in the prompt.
   • CIncrease the number of tokens in the prompt.
   • DUse Provisioned Throughput.

   Show answer

   Answer: B

   Explanation

   Bedrock pricing is based on the number of tokens processed, which includes both input tokens (from the prompt) and output tokens (generated by the model). By decreasing the number of tokens in the prompt, you directly reduce the cost associated with each invocation of the model.

   Source: Adapted from community discussion — Blair77

   Next question

6. Question 126 · 417 in bank

   An AI practitioner is using a large language model (LLM) to create content for marketing campaigns. The generated content sounds plausible and factual but is incorrect. Which problem is the LLM having?

   • AData leakage
   • BHallucination
   • COverfitting
   • DUnderfitting

   Show answer

   Answer: B

   Explanation

   Hallucination is a phenomenon in which an LLM generates text that sounds plausible and factual but is actually incorrect or nonsensical. This occurs when the model is overconfident in its ability to generate coherent text based on patterns it has learned from training data.

   Source: Adapted from community discussion — AzureDP900

   Next question

7. Question 127 · 417 in bank

   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?

   • ADelete the custom model. Remove the confidential data from the training dataset. Retrain the custom model.
   • BMask the confidential data in the inference responses by using dynamic data masking.
   • CEncrypt the confidential data in the inference responses by using Amazon SageMaker.
   • DEncrypt the confidential data in the custom model by using AWS Key Management Service (AWS KMS).

   Show answer

   Answer: A

   Explanation

   Once a model is trained on data, its outputs may inherently reflect patterns or details derived from the training dataset, including confidential data. To ensure the custom model does not generate inference responses based on confidential data, the only reliable solution is to: - Remove the confidential data from the training dataset. - Retrain the model with the updated dataset. This approach ensures the model is not influenced by sensitive information during inference. Option B is incorrect. Dynamic data masking hides sensitive information in database query results or outputs but does not prevent the model from generating responses influenced by the confidential data. The model would still "know" the sensitive patterns.

   Source: Adapted from community discussion — ap6491

   Next question

8. Question 128 · 417 in bank

   A company has built a solution by using generative AI. The solution uses large language models (LLMs) to translate training manuals from English into other languages. The company wants to evaluate the accuracy of the solution by examining the text generated for the manuals. Which model evaluation strategy meets these requirements?

   • ABilingual Evaluation Understudy (BLEU)
   • BRoot mean squared error (RMSE)
   • CRecall-Oriented Understudy for Gisting Evaluation (ROUGE)
   • DF1 score

   Show answer

   Answer: A

   Explanation

   BLEU (bilingual evaluation understudy) is an algorithm for evaluating the quality of text which has been machine-translated from one natural language to another.

   Source: Adapted from community discussion — Amitst

   Next question

9. Question 129 · 417 in bank

   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?

   • AHuman-in-the-loop validation by using Amazon SageMaker Ground Truth Plus
   • BData augmentation by using an Amazon Bedrock knowledge base
   • CImage recognition by using Amazon Rekognition
   • DData summarization by using Amazon QuickSight Q

   Show answer

   Answer: A

   Explanation

   Human-in-the-loop validation combines the efficiency of machine learning with human expertise to ensure high-quality labeled data. Amazon SageMaker Ground Truth Plus enables you to have human labelers validate and correct model predictions, which reduces errors in annotations and increases the accuracy of the training data. This is particularly useful when you need to generate accurate images or annotations for protective eyewear and want to ensure that the annotations are reliable.

   Source: Adapted from community discussion — Jessiii

   References

   • AWS documentation

   Next question

10. Question 130 · 417 in bank

    A large retailer receives thousands of customer support inquiries about products every day. The customer support inquiries need to be processed and responded to quickly. The company wants to implement Agents for Amazon Bedrock. What are the key benefits of using Amazon Bedrock agents that could help this retailer?

    • AGeneration of custom foundation models (FMs) to predict customer needs
    • BAutomation of repetitive tasks and orchestration of complex workflows
    • CAutomatically calling multiple foundation models (FMs) and consolidating the results
    • DSelecting the foundation model (FM) based on predefined criteria and metrics

    Show answer

    Answer: B

    Explanation

    B. Main advantage of using Amazon Bedrock Agent is the automation of repetitive tasks and the orchestration of complex workflows. Customer support inquiries are often patterned and repetitive. By using Amazon Bedrock Agent, you can automate the initial response to such routine inquiries and information gathering. In addition, even in cases where the workflow is somewhat complex and the response required varies depending on the inquiry content, Amazon Bedrock can flexibly respond by combining multiple AI models. This will enable quick and accurate customer support while optimally utilizing human resources. It will lead to efficient processing of large volumes of inquiries, which will lead to improved customer satisfaction.

    Source: Adapted from community discussion — taka5094

    References

    • AWS documentation

    Next question

Keyboard:J Kmove between questionsA–Dselect an answerRshow or hide the answer← →change page

AIF-C01 practice, question 121. 417 questions available.