AWS Certified AI Practitioner AIF-C01 Practice Test 4
Question 30
A data analyst receives transaction tables, customer chat text and store-camera photos for one research initiative. Which option BEST meets the business requirement?
- Treat rows as structured tabular data and text/images as unstructured modalities
- Assume labeled and unlabeled mean the same thing as structured and unstructured
- Treat every numeric transaction column as a video frame
- Classify every input as tabular merely because it is stored in S3
Explanation
Correct Answer
A:Each modality has different representation and preprocessing requirements The specified scenario makes this the most appropriate approach.
Incorrect Answers
B:Target-label availability and data structure are separate properties This prevents the approach from meeting the stated requirement.
C:Numeric table columns are not image or video pixels This prevents the approach from meeting the stated requirement.
D:Object-storage location does not make image pixels or natural text tabular This prevents the approach from meeting the stated requirement.
Question 31
A streaming company has no audience-category labels but wants to discover patterns in viewing behavior. After selecting the recommended approach, which evidence MOST DIRECTLY demonstrates that the actual required outcome has been achieved?
- Inspect separation and business relevance of clusters on unseen user behavior data
- A component test invokes this alternate approach successfully: Use time-series forecasting to produce only next week's total traffic
- A reviewed design record approves this competing approach: Use translation of descriptions as the segmentation algorithm
- A system inventory lists this competing approach as active: Train supervised classification using nonexistent segment labels
Explanation
Correct Answer
A:Inspect separation and business relevance of clusters on unseen user behavior data. This checks the business result rather than just a configured feature.
Incorrect Answers
B:Forecasting aggregates volume rather than discovering user groups A component invocation does not validate the necessary end-to-end outcome.
C:Language translation does not group viewing patterns by behavior Design approval is not evidence that the alternative meets the actual requirement.
D:A supervised classifier requires an appropriate labeled target An active setting cannot independently prove the requested business outcome.
Question 32
A product manager confuses generative AI with ordinary binary prediction of account fraud. A pilot uses this approach instead: Treat every neural network's output as necessarily a generated paragraph. Which limitation MOST DIRECTLY explains why the requirement is not met?
- Neural networks can be used for classification and regression as well as generation
- A binary fraud label is a prediction task; generation produces new text or other content
- A tabular classifier may be sufficient and cheaper for labeled fixed classes
- Deep learning is a family of machine learning methods
Explanation
Correct Answer
A:This is the actual failure mode of the proposed approach: Treat every neural network's output as necessarily a generated paragraph. Neural networks can be used for classification and regression as well as generation
Incorrect Answers
B:A binary fraud label is a prediction task; generation produces new text or other content This explains the recommended design, but it is not the requested diagnosis of Treat every neural network's output as necessarily a generated paragraph.
C:This limitation concerns Assume a language model is required for every tabular classification; it does not diagnose the proposed Treat every neural network's output as necessarily a generated paragraph approach.
D:This limitation concerns Describe deep learning as unrelated to machine learning; it does not diagnose the proposed Treat every neural network's output as necessarily a generated paragraph approach.
Question 33
An accounts-payable department needs to extract invoice fields and table cells from scanned PDF documents. Before approving a rollout, an architect suggests Use Amazon Polly to read the scanned PDF as an invoice database. Which mismatch should be documented for this precise design?
- Textract is designed to extract text and structured fields or tables from document images
- Visual object labels do not reconstruct invoice form keys and table cell structure
- Speech synthesis does not parse scanned invoice forms
- Language translation does not identify document tables and fields
Explanation
Correct Answer
C:This is the actual failure mode of the proposed approach: Use Amazon Polly to read the scanned PDF as an invoice database. Speech synthesis does not parse scanned invoice forms
Incorrect Answers
A:Textract is designed to extract text and structured fields or tables from document images This explains the recommended design, but it is not the requested diagnosis of Use Amazon Polly to read the scanned PDF as an invoice database.
B:This limitation concerns Use Amazon Rekognition general object labels as the invoice parser; it does not diagnose the proposed Use Amazon Polly to read the scanned PDF as an invoice database approach.
D:This limitation concerns Use Amazon Translate to extract source PDF table coordinates; it does not diagnose the proposed Use Amazon Polly to read the scanned PDF as an invoice database approach.
Question 34
A transit application needs to read an already generated arrival announcement aloud for accessibility. During a design review, someone proposes Use Amazon Translate as the text-to-speech engine. Which technical objection is specifically relevant to this proposal?
- Transcribe converts speech to text in the opposite direction
- Recognizing entities does not render speech
- Translate produces translated text rather than spoken synthesis
- Polly creates spoken audio from supplied text
Explanation
Correct Answer
C:This is the actual failure mode of the proposed approach: Use Amazon Translate as the text-to-speech engine. Translate produces translated text rather than spoken synthesis
Incorrect Answers
A:This limitation concerns Use Amazon Transcribe to turn the arrival text into an audio clip; it does not diagnose the proposed Use Amazon Translate as the text-to-speech engine approach.
B:This limitation concerns Use Amazon Comprehend entity detection as the speech generator; it does not diagnose the proposed Use Amazon Translate as the text-to-speech engine approach.
D:Polly creates spoken audio from supplied text This explains the recommended design, but it is not the requested diagnosis of Use Amazon Translate as the text-to-speech engine.
Question 35
An analytics team needs a managed way to experiment with and deploy its own specialized forecasting ML model. Before approving a rollout, an architect suggests Use Amazon Translate as the experiment tracking and model hosting service. Which mismatch should be documented for this precise design?
- Speech synthesis is not a general ML training and deployment platform
- Translation does not supply a custom ML development lifecycle
- SageMaker AI supports developing, training, evaluating and hosting custom ML workloads
- FM inference alone does not train the specified bespoke forecasting pipeline
Explanation
Correct Answer
B:This is the actual failure mode of the proposed approach: Use Amazon Translate as the experiment tracking and model hosting service. Translation does not supply a custom ML development lifecycle
Incorrect Answers
A:This limitation concerns Use Amazon Polly to train the demand forecasting model; it does not diagnose the proposed Use Amazon Translate as the experiment tracking and model hosting service approach.
C:SageMaker AI supports developing, training, evaluating and hosting custom ML workloads This explains the recommended design, but it is not the requested diagnosis of Use Amazon Translate as the experiment tracking and model hosting service.
D:This limitation concerns Treat an Amazon Bedrock managed FM invocation as a custom regression-training job; it does not diagnose the proposed Use Amazon Translate as the experiment tracking and model hosting service approach.
Question 36
A data analyst receives transaction tables, customer chat text and store-camera photos for one research initiative. During a design review, someone proposes Treat every numeric transaction column as a video frame. Which technical objection is specifically relevant to this proposal?
- Numeric table columns are not image or video pixels
- Object-storage location does not make image pixels or natural text tabular
- Each modality has different representation and preprocessing requirements
- Target-label availability and data structure are separate properties
Explanation
Correct Answer
A:This is the actual failure mode of the proposed approach: Treat every numeric transaction column as a video frame. Numeric table columns are not image or video pixels
Incorrect Answers
B:This limitation concerns Classify every input as tabular merely because it is stored in S3; it does not diagnose the proposed Treat every numeric transaction column as a video frame approach.
C:Each modality has different representation and preprocessing requirements This explains the recommended design, but it is not the requested diagnosis of Treat every numeric transaction column as a video frame.
D:This limitation concerns Assume labeled and unlabeled mean the same thing as structured and unstructured; it does not diagnose the proposed Treat every numeric transaction column as a video frame approach.
Question 37
A simulated warehouse robot learns which route to choose from delayed rewards for delivery time and collisions. A pilot uses this approach instead: Use unsupervised clustering to assign static product categories. Which limitation MOST DIRECTLY explains why the requirement is not met?
- The agent learns actions from rewards and environmental feedback across episodes
- A price estimate does not determine route actions over time
- Clustering does not optimize sequences of robot actions against rewards
- Recognizing labels is not a learned navigation policy driven by rewards
Explanation
Correct Answer
C:This is the actual failure mode of the proposed approach: Use unsupervised clustering to assign static product categories. Clustering does not optimize sequences of robot actions against rewards
Incorrect Answers
A:The agent learns actions from rewards and environmental feedback across episodes This explains the recommended design, but it is not the requested diagnosis of Use unsupervised clustering to assign static product categories.
B:This limitation concerns Use standard regression to output a single shipping price; it does not diagnose the proposed Use unsupervised clustering to assign static product categories approach.
D:This limitation concerns Use supervised OCR to read every shelf label as the entire control policy; it does not diagnose the proposed Use unsupervised clustering to assign static product categories approach.
Question 38
A data analyst receives transaction tables, customer chat text and store-camera photos for one research initiative. Before approving a rollout, an architect suggests Assume labeled and unlabeled mean the same thing as structured and unstructured. Which mismatch should be documented for this precise design?
- Object-storage location does not make image pixels or natural text tabular
- Numeric table columns are not image or video pixels
- Each modality has different representation and preprocessing requirements
- Target-label availability and data structure are separate properties
Explanation
Correct Answer
D:This is the actual failure mode of the proposed approach: Assume labeled and unlabeled mean the same thing as structured and unstructured. Target-label availability and data structure are separate properties
Incorrect Answers
A:This limitation concerns Classify every input as tabular merely because it is stored in S3; it does not diagnose the proposed Assume labeled and unlabeled mean the same thing as structured and unstructured approach.
B:This limitation concerns Treat every numeric transaction column as a video frame; it does not diagnose the proposed Assume labeled and unlabeled mean the same thing as structured and unstructured approach.
C:Each modality has different representation and preprocessing requirements This explains the recommended design, but it is not the requested diagnosis of Assume labeled and unlabeled mean the same thing as structured and unstructured.
Question 39
A binary disease-screening model misses many true positive cases although its overall accuracy looks high. 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 mean absolute error as the primary binary confusion-matrix statistic
- The confusion matrix confirms improved true-positive capture without an unacceptable precision tradeoff
- A reviewed design record approves this competing approach: Use token output cost as the medical false-negative measure
- A system inventory lists this competing approach as active: Use only accuracy without inspecting class prevalence
Explanation
Correct Answer
B:The confusion matrix confirms improved true-positive capture without an unacceptable precision tradeoff. This checks the business result rather than just a configured feature.
Incorrect Answers
A:A continuous regression metric does not directly show missed positive classifications A component invocation does not validate the necessary end-to-end outcome.
C:Inference spending does not count clinical positives missed Design approval is not evidence that the alternative meets the actual requirement.
D:Accuracy can appear high when negative cases dominate An active setting cannot independently prove the requested business outcome.
Question 40
A model training project fails because currency fields contain strings and missing values are inconsistent. During a design review, someone proposes Skip training-data validation and rely only on more expensive compute. Which technical objection is specifically relevant to this proposal?
- Extra compute cannot guarantee consistent features or valid numeric inputs
- Sampling temperature does not repair tabular training data quality
- Preprocessing resolves malformed values and missing-data treatment that otherwise corrupt features
- Monitoring after deployment does not clean malformed training records
Explanation
Correct Answer
A:This is the actual failure mode of the proposed approach: Skip training-data validation and rely only on more expensive compute. Extra compute cannot guarantee consistent features or valid numeric inputs
Incorrect Answers
B:This limitation concerns Tune output temperature on an existing generative model to repair a broken CSV; it does not diagnose the proposed Skip training-data validation and rely only on more expensive compute approach.
C: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 Skip training-data validation and rely only on more expensive compute.
D:This limitation concerns Assume deployment monitoring replaces initial data preparation; it does not diagnose the proposed Skip training-data validation and rely only on more expensive compute approach.
