McAfee Secure

Exam Code: AI-901

Exam Name: Microsoft Azure AI Fundamentals

Certification Provider: Microsoft

Corresponding Certification: Microsoft Certified: Azure AI Fundamentals

Microsoft AI-901 Questions & Answers

Study with Up-To-Date REAL Exam Questions and Answers from the ACTUAL Test

166 Questions & Answers with Testing Engine

"Microsoft Azure AI Fundamentals Exam", also known as AI-901 exam, is a Microsoft certification exam.

Pass your tests with the always up-to-date AI-901 Exam Engine. Your AI-901 training materials keep you at the head of the pack!

Satisfaction Guaranteed

Money Back Guarantee

Testking provides hassle-free money back guarantee with our products. That is because we have 100% trust in the abilities of our professional and experience product team, and our record is a proof of that.

99.6% PASS RATE
Was: $137.49
Now: $124.99

Microsoft AI-901 Practice Test 62

Question 610

Troubleshooting report: Overlapping detections are counted twice. Which correction is MOST directly targeted to this failure?

  1. Gate distribution on the image moderation result
  2. Add confidence gating and supporting evidence
  3. Apply confidence and nonmaximum suppression or instance deduplication
  4. Route to the text-to-image mode with intended inputs only

Explanation

Correct Answer

C:

This correction directly addresses the reported failure. Object counts require resolving distinct instances, not trusting one scene label. The targeted change is: Apply confidence and nonmaximum suppression or instance deduplication.

Incorrect Answers

A:

Generative media needs policy controls in addition to prompt instructions. In the stated case, troubleshooting report: Overlapping detections are counted twice; “Gate distribution on the image moderation result” does not target that failure.

B:

Generated visual interpretations can be wrong even when phrased fluently. In the stated case, troubleshooting report: Overlapping detections are counted twice; “Add confidence gating and supporting evidence” does not target that failure.

D:

Generation from text is distinct from image editing. In the stated case, troubleshooting report: Overlapping detections are counted twice; “Route to the text-to-image mode with intended inputs only” does not target that failure.

Question 611

An engineering review documents this requirement: An application must extract text from a photographed license plate. Select the most appropriate solution.

  1. Use image text recognition with location evidence
  2. Use layout-aware document analysis that preserves rows, selection marks and text spans
  3. Enable supported document-field confidence and grounding, then calibrate field-specific review thresholds using labeled samples
  4. Use a video-capable Content Understanding analyzer with timestamps and supported fields

Explanation

Correct Answer

A:

OCR is appropriate for characters embedded in a photograph. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

B:

Ordinary plain-text OCR may lose table relationships and checkbox states. That is not sufficient here: OCR is appropriate for characters embedded in a photograph.

C:

Content Understanding confidence applies to supported document analyzers; confidence estimates alone do not prove field correctness. That is not sufficient here: OCR is appropriate for characters embedded in a photograph.

D:

Multimodal extraction draws on visual and spoken evidence over time. That is not sufficient here: OCR is appropriate for characters embedded in a photograph.

Question 612

An Azure AI solution uses the following approved control: Check playback device selection and the synthesized audio format. What observable result should pass QA?

  1. Mixed test calls identify the correct language before transcription
  2. The transcript matches a labeled spoken test phrase
  3. Mic capture ceases and the app no longer receives audio frames
  4. The generated audio plays through the selected output device

Explanation

Correct Answer

D:

Successful synthesis does not guarantee local audio is routed to a speaker. The expected acceptance evidence is: The generated audio plays through the selected output device. This observation directly tests the stated behavior.

Incorrect Answers

A:

Automatic language detection chooses among an approved set of languages. This observation does not establish the required result: The generated audio plays through the selected output device.

B:

Speech recognition converts waveform features to text tokens. This observation does not establish the required result: The generated audio plays through the selected output device.

C:

User control of capture must be enforced at the device or application layer. This observation does not establish the required result: The generated audio plays through the selected output device.

Question 613

An Azure AI Fundamentals solution has this requirement: A vision result reports a bounding box for each pedestrian. Which implementation most directly addresses it?

  1. Provide the actual image to a capable multimodal or OCR pipeline
  2. Use the box coordinates to locate detected people in the source image
  3. Preserve region coordinates with extracted text or field evidence
  4. Use an image-generation-capable model with an appropriate prompt

Explanation

Correct Answer

B:

Localization metadata is tied to pixel regions, not just labels. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

Visual evidence is needed to read sign content. That is not sufficient here: Localization metadata is tied to pixel regions, not just labels.

C:

Spatial provenance supports later verification of recognition results. That is not sufficient here: Localization metadata is tied to pixel regions, not just labels.

D:

Image generation creates new pixels rather than classifying an existing photo. That is not sufficient here: Localization metadata is tied to pixel regions, not just labels.

Question 614

An Azure AI solution uses the following approved control: Confirm the request includes actual supported image content and an image-capable deployment. What observable result should pass QA?

  1. The test workload recovers without unbounded retry bursts
  2. A unique feature in the image appears correctly in the response
  3. A correctly converted test image processes successfully
  4. Prominent objects receive separate region captions and coordinates

Explanation

Correct Answer

B:

A textual filename cannot convey pixels to a model. The expected acceptance evidence is: A unique feature in the image appears correctly in the response. This observation directly tests the stated behavior.

Incorrect Answers

A:

Too many concurrent requests can exceed resource quota. This observation does not establish the required result: A unique feature in the image appears correctly in the response.

C:

Vision endpoints impose supported input formats and limits. This observation does not establish the required result: A unique feature in the image appears correctly in the response.

D:

Dense captions describe localized regions beyond a single global caption. This observation does not establish the required result: A unique feature in the image appears correctly in the response.

Question 615

Troubleshooting report: The workflow assumes a single unreviewed output always matches branding. Which correction is MOST directly targeted to this failure?

  1. Use the documented coordinate units and image dimensions
  2. Apply confidence and nonmaximum suppression or instance deduplication
  3. Constrain the prompt and validate the final artwork
  4. Collect balanced labeled examples and test class confusion

Explanation

Correct Answer

C:

This correction directly addresses the reported failure. Explicit visual guidance can influence generated appearance but is not absolute. The targeted change is: Constrain the prompt and validate the final artwork.

Incorrect Answers

A:

Localization metadata is tied to pixel regions, not just labels. In the stated case, troubleshooting report: The workflow assumes a single unreviewed output always matches branding; “Use the documented coordinate units and image dimensions” does not target that failure.

B:

Object counts require resolving distinct instances, not trusting one scene label. In the stated case, troubleshooting report: The workflow assumes a single unreviewed output always matches branding; “Apply confidence and nonmaximum suppression or instance deduplication” does not target that failure.

D:

Fine-grained visual classes require representative discriminating images. In the stated case, troubleshooting report: The workflow assumes a single unreviewed output always matches branding; “Collect balanced labeled examples and test class confusion” does not target that failure.

Question 616

Troubleshooting report: The model fabricates a purchase-order number. Which correction is MOST directly targeted to this failure?

  1. Keep temporal and content provenance in the retrieval result
  2. Validate field presence against source evidence
  3. Use layout-aware field definitions rather than fixed crop positions
  4. Use schema-aware field disambiguation and source labels

Explanation

Correct Answer

B:

This correction directly addresses the reported failure. Extraction should preserve absence and distinguish it from zero. The targeted change is: Validate field presence against source evidence.

Incorrect Answers

A:

Multimodal grounding should preserve how the answer maps to source media. In the stated case, troubleshooting report: The model fabricates a purchase-order number; “Keep temporal and content provenance in the retrieval result” does not target that failure.

C:

A schema-driven analyzer identifies business fields independent of fixed coordinates. In the stated case, troubleshooting report: The model fabricates a purchase-order number; “Use layout-aware field definitions rather than fixed crop positions” does not target that failure.

D:

Typed extraction avoids mixing semantically different dates. In the stated case, troubleshooting report: The model fabricates a purchase-order number; “Use schema-aware field disambiguation and source labels” does not target that failure.

Question 617

An engineering review documents this requirement: A model supplies a fluent answer to a question outside its training evidence. Select the most appropriate solution.

  1. Represent text with embeddings and compare vectors using an appropriate similarity measure
  2. Require verifiable grounding and calibrated uncertainty
  3. Use a supervised classification model with representative labeled cases
  4. Use clustering or another appropriate unsupervised technique

Explanation

Correct Answer

B:

Language-model likelihood does not independently establish factual truth. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

Embeddings encode learned semantic relationships rather than literal string identity. That is not sufficient here: Language-model likelihood does not independently establish factual truth.

C:

Classification predicts discrete labels rather than a continuous numeric target. That is not sufficient here: Language-model likelihood does not independently establish factual truth.

D:

Unsupervised learning finds structure without supplied target labels. That is not sufficient here: Language-model likelihood does not independently establish factual truth.

Question 618

An Azure-based AI team must satisfy this objective: A model handles an unusually ambiguous natural-language request as a definitive instruction. Which design choice is best?

  1. Include rare but consequential safety edge cases in acceptance tests
  2. Ask for clarification or abstain rather than acting on inferred intent
  3. Evaluate instructions and fallback quality in each supported language
  4. Require governance evidence with reviewers, test outcomes, and documented exceptions

Explanation

Correct Answer

B:

Reliable assistance distinguishes missing intent from authorized action. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

A high overall score can conceal severe failure on unusual inputs. That is not sufficient here: Reliable assistance distinguishes missing intent from authorized action.

C:

Inclusion requires equivalent safety and meaning across supported locales. That is not sufficient here: Reliable assistance distinguishes missing intent from authorized action.

D:

Self-attestation is not independent proof of regulatory controls. That is not sufficient here: Reliable assistance distinguishes missing intent from authorized action.

Question 619

An Azure AI solution uses the following approved control: Use optical character recognition instead of generic image tagging. What observable result should pass QA?

  1. The model/deployment capability check and a spoken-audio test confirm that this endpoint supports transcription
  2. The resulting cohorts are evaluated for cohesion and business usefulness
  3. The returned text includes the photographed serial number
  4. The response includes the requested image artifact and metadata

Explanation

Correct Answer

C:

OCR produces readable text from pixels. The expected acceptance evidence is: The returned text includes the photographed serial number. This observation directly tests the stated behavior.

Incorrect Answers

A:

Model capabilities must include the required audio modality before the client can obtain speech-to-text results. This observation does not establish the required result: The returned text includes the photographed serial number.

B:

Unsupervised learning finds structure without supplied target labels. This observation does not establish the required result: The returned text includes the photographed serial number.

D:

Text generation and image generation are distinct model output capabilities. This observation does not establish the required result: The returned text includes the photographed serial number.

Question 620

Troubleshooting report: Evaluation uses only studio-recorded standard-accent speech. Which correction is MOST directly targeted to this failure?

  1. Enforce backend authorization and signed approval regardless of prompt wording
  2. Expand the consented voice test corpus and provide fallback interaction
  3. Add low-light and motion-blur scenarios to the evaluation set
  4. Propagate typed tool errors and use an explicit unavailable response

Explanation

Correct Answer

B:

This correction directly addresses the reported failure. Inclusive voice systems require performance evaluation beyond a single accent. The targeted change is: Expand the consented voice test corpus and provide fallback interaction.

Incorrect Answers

A:

Accountable high-impact operations need verifiable authorized signatures. In the stated case, troubleshooting report: Evaluation uses only studio-recorded standard-accent speech; “Enforce backend authorization and signed approval regardless of prompt wording” does not target that failure.

C:

Reliability demands testing the actual operating environment, not just clean samples. In the stated case, troubleshooting report: Evaluation uses only studio-recorded standard-accent speech; “Add low-light and motion-blur scenarios to the evaluation set” does not target that failure.

D:

A fluent generation cannot substitute for missing authoritative transaction data. In the stated case, troubleshooting report: Evaluation uses only studio-recorded standard-accent speech; “Propagate typed tool errors and use an explicit unavailable response” does not target that failure.

Microsoft Azure AI Fundamentals (AI-901) Practice Tests