AWS Certified AI Practitioner AIF-C01 Practice Test 7
Question 60
An insurance provider receives recordings of approved claims interviews and needs searchable text transcripts. After selecting the recommended approach, which evidence MOST DIRECTLY demonstrates that the actual required outcome has been achieved?
- A reviewed design record approves this competing approach: Use Amazon Rekognition labels to extract spoken words from audio
- A system inventory lists this competing approach as active: Use Amazon Polly for the audio-to-text conversion
- A component test invokes this alternate approach successfully: Use Amazon Translate without first obtaining a transcript
- Compare time-aligned transcripts with representative reviewed recordings
Explanation
Correct Answer
D:Compare time-aligned transcripts with representative reviewed recordings. This checks the business result rather than just a configured feature.
Incorrect Answers
A:Rekognition image/video analysis is not a general speech transcription service Design approval is not evidence that the alternative meets the actual requirement.
B:Polly generates audio from text rather than transcribing speech An active setting cannot independently prove the requested business outcome.
C:Translation alone does not perform the requested speech recognition A component invocation does not validate the necessary end-to-end outcome.
Question 61
An operations team must process a million stored images overnight rather than respond to user uploads instantly. A pilot uses this approach instead: Choose only synchronous low-latency requests for each image. Which limitation MOST DIRECTLY explains why the requirement is not met?
- Bulk offline prediction need not maintain low per-request interactive latency
- Re-training per input is not normal inference and adds avoidable cost
- Audio conversion does not produce requested visual predictions
- Interactive calls add unnecessary per-request pressure when offline processing is allowed
Explanation
Correct Answer
D:This is the actual failure mode of the proposed approach: Choose only synchronous low-latency requests for each image. Interactive calls add unnecessary per-request pressure when offline processing is allowed
Incorrect Answers
A:Bulk offline prediction need not maintain low per-request interactive latency This explains the recommended design, but it is not the requested diagnosis of Choose only synchronous low-latency requests for each image.
B:This limitation concerns Train a new model for each uploaded image; it does not diagnose the proposed Choose only synchronous low-latency requests for each image approach.
C:This limitation concerns Replace image inference with an audio conversion service; it does not diagnose the proposed Choose only synchronous low-latency requests for each image approach.
Question 62
A model training project fails because currency fields contain strings and missing values are inconsistent. A pilot uses this approach instead: Tune output temperature on an existing generative model to repair a broken CSV. Which limitation MOST DIRECTLY explains why the requirement is not met?
- Extra compute cannot guarantee consistent features or valid numeric inputs
- Preprocessing resolves malformed values and missing-data treatment that otherwise corrupt features
- Sampling temperature does not repair tabular training data quality
- Monitoring after deployment does not clean malformed training records
Explanation
Correct Answer
C:This is the actual failure mode of the proposed approach: Tune output temperature on an existing generative model to repair a broken CSV. Sampling temperature does not repair tabular training data quality
Incorrect Answers
A:This limitation concerns Skip training-data validation and rely only on more expensive compute; it does not diagnose the proposed Tune output temperature on an existing generative model to repair a broken CSV approach.
B:Preprocessing resolves malformed values and missing-data treatment that otherwise corrupt features This explains the recommended design, but it is not the requested diagnosis of Tune output temperature on an existing generative model to repair a broken CSV.
D:This limitation concerns Assume deployment monitoring replaces initial data preparation; it does not diagnose the proposed Tune output temperature on an existing generative model to repair a broken CSV approach.
Question 63
A hotel wants a conversational bot that collects check-in date and room type through voice or text turns. Which option BEST meets the business requirement?
- Use Amazon Rekognition to infer date and room type from speech
- Use Amazon Polly alone as the complete booking dialog engine
- Use Amazon Translate alone to track hotel reservation state
- Use Amazon Lex intents and slots for a task-oriented conversational interface
Explanation
Correct Answer
D:Lex is designed for conversational intent recognition and slot elicitation The specified scenario makes this the most appropriate approach.
Incorrect Answers
A:Image analysis is not a spoken reservation dialog manager This prevents the approach from meeting the stated requirement.
B:Polly speaks text but does not manage user intents or slot collection This prevents the approach from meeting the stated requirement.
C:Translation does not orchestrate multi-turn reservation fields This prevents the approach from meeting the stated requirement.
Question 64
An online checkout system must return a fraud score while the customer is still waiting. After selecting the recommended approach, which evidence MOST DIRECTLY demonstrates that the actual required outcome has been achieved?
- A system inventory lists this competing approach as active: Schedule a single nightly batch job for all checkout requests
- A reviewed design record approves this competing approach: Use a translation pipeline as the fraud scoring endpoint
- P95 inference latency and prediction availability meet the defined checkout service-level target
- A component test invokes this alternate approach successfully: Require humans to retrain the model before every purchase
Explanation
Correct Answer
C:P95 inference latency and prediction availability meet the defined checkout service-level target. This checks the business result rather than just a configured feature.
Incorrect Answers
A:Nightly batch results arrive after the checkout decision window An active setting cannot independently prove the requested business outcome.
B:Translating purchase descriptions is not a fraud score Design approval is not evidence that the alternative meets the actual requirement.
D:Model training is separate from serving each inference request A component invocation does not validate the necessary end-to-end outcome.
Question 65
An email service has thousands of messages labeled spam or legitimate and needs to classify future messages. After selecting the recommended approach, which evidence MOST DIRECTLY demonstrates that the actual required outcome has been achieved?
- A component test invokes this alternate approach successfully: Use regression to predict a continuous dollar amount
- A system inventory lists this competing approach as active: Use unsupervised clustering without consulting existing labels
- A held-out set reports spam precision and recall against verified message labels
- A reviewed design record approves this competing approach: Use reinforcement learning with rewards for warehouse robots
Explanation
Correct Answer
C:A held-out set reports spam precision and recall against verified message labels. This checks the business result rather than just a configured feature.
Incorrect Answers
A:Regression predicts numerical quantities rather than the specified discrete spam class A component invocation does not validate the necessary end-to-end outcome.
B:Clustering groups by similarity without directly learning the recorded spam labels An active setting cannot independently prove the requested business outcome.
D:A reward-driven control policy does not directly exploit the provided labeled email classes Design approval is not evidence that the alternative meets the actual requirement.
Question 66
A transit application needs to read an already generated arrival announcement aloud for accessibility. Which option BEST meets the business requirement?
- Use Amazon Polly text-to-speech synthesis
- Use Amazon Translate as the text-to-speech engine
- Use Amazon Comprehend entity detection as the speech generator
- Use Amazon Transcribe to turn the arrival text into an audio clip
Explanation
Correct Answer
A:Polly creates spoken audio from supplied text The specified scenario makes this the most appropriate approach.
Incorrect Answers
B:Translate produces translated text rather than spoken synthesis This prevents the approach from meeting the stated requirement.
C:Recognizing entities does not render speech This prevents the approach from meeting the stated requirement.
D:Transcribe converts speech to text in the opposite direction This prevents the approach from meeting the stated requirement.
Question 67
An analytics team needs a managed way to experiment with and deploy its own specialized forecasting ML model. After selecting the recommended approach, which evidence MOST DIRECTLY demonstrates that the actual required outcome has been achieved?
- A reviewed design record approves this competing approach: Use Amazon Translate as the experiment tracking and model hosting service
- A system inventory lists this competing approach as active: Treat an Amazon Bedrock managed FM invocation as a custom regression-training job
- A component test invokes this alternate approach successfully: Use Amazon Polly to train the demand forecasting model
- A registered custom forecasting model is evaluated on held-out data and invoked through the intended serving path
Explanation
Correct Answer
D:A registered custom forecasting model is evaluated on held-out data and invoked through the intended serving path. This checks the business result rather than just a configured feature.
Incorrect Answers
A:Translation does not supply a custom ML development lifecycle Design approval is not evidence that the alternative meets the actual requirement.
B:FM inference alone does not train the specified bespoke forecasting pipeline An active setting cannot independently prove the requested business outcome.
C:Speech synthesis is not a general ML training and deployment platform A component invocation does not validate the necessary end-to-end outcome.
Question 68
A product manager confuses generative AI with ordinary binary prediction of account fraud. Which option BEST meets the business requirement?
- Distinguish predictive classification from generative content creation
- Assume a language model is required for every tabular classification
- Describe deep learning as unrelated to machine learning
- Treat every neural network's output as necessarily a generated paragraph
Explanation
Correct Answer
A:A binary fraud label is a prediction task; generation produces new text or other content The specified scenario makes this the most appropriate approach.
Incorrect Answers
B:A tabular classifier may be sufficient and cheaper for labeled fixed classes This prevents the approach from meeting the stated requirement.
C:Deep learning is a family of machine learning methods This prevents the approach from meeting the stated requirement.
D:Neural networks can be used for classification and regression as well as generation This prevents the approach from meeting the stated requirement.
Question 69
An email service has thousands of messages labeled spam or legitimate and needs to classify future messages. During a design review, someone proposes Use regression to predict a continuous dollar amount. Which technical objection is specifically relevant to this proposal?
- Clustering groups by similarity without directly learning the recorded spam labels
- Regression predicts numerical quantities rather than the specified discrete spam class
- Labels identify the target classes and allow the classifier to learn a decision boundary
- A reward-driven control policy does not directly exploit the provided labeled email classes
Explanation
Correct Answer
B:This is the actual failure mode of the proposed approach: Use regression to predict a continuous dollar amount. Regression predicts numerical quantities rather than the specified discrete spam class
Incorrect Answers
A:This limitation concerns Use unsupervised clustering without consulting existing labels; it does not diagnose the proposed Use regression to predict a continuous dollar amount approach.
C:Labels identify the target classes and allow the classifier to learn a decision boundary This explains the recommended design, but it is not the requested diagnosis of Use regression to predict a continuous dollar amount.
D:This limitation concerns Use reinforcement learning with rewards for warehouse robots; it does not diagnose the proposed Use regression to predict a continuous dollar amount approach.
Question 70
An insurance provider receives recordings of approved claims interviews and needs searchable text transcripts. Before approving a rollout, an architect suggests Use Amazon Rekognition labels to extract spoken words from audio. Which mismatch should be documented for this precise design?
- Automatic speech recognition converts audio speech into written transcript content
- Rekognition image/video analysis is not a general speech transcription service
- Translation alone does not perform the requested speech recognition
- Polly generates audio from text rather than transcribing speech
Explanation
Correct Answer
B:This is the actual failure mode of the proposed approach: Use Amazon Rekognition labels to extract spoken words from audio. Rekognition image/video analysis is not a general speech transcription service
Incorrect Answers
A:Automatic speech recognition converts audio speech into written transcript content This explains the recommended design, but it is not the requested diagnosis of Use Amazon Rekognition labels to extract spoken words from audio.
C:This limitation concerns Use Amazon Translate without first obtaining a transcript; it does not diagnose the proposed Use Amazon Rekognition labels to extract spoken words from audio approach.
D:This limitation concerns Use Amazon Polly for the audio-to-text conversion; it does not diagnose the proposed Use Amazon Rekognition labels to extract spoken words from audio approach.
