A data scientist is labeling images for a custom image classification model that will distinguish between cats and dogs. What is the correct labeling approach?

Tag each image with both 'cat' and 'dog' regardless of content

Assign the tag 'cat' or 'dog' to each entire image

Draw a bounding box around each cat and dog and tag each box

Provide a text file describing the animals in each image

An application is calling a published custom vision prediction endpoint. The response includes multiple tags with probabilities. The application needs to decide whether to act on a prediction. What should it do?

Use the tag with the highest probability and compare it to a business-defined threshold

Always act on the tag with the highest probability regardless of value

Act on all tags with probability above 0.5

Ignore the probabilities and use the first tag in the response

What is the purpose of the 'Training iterations' in Azure Custom Vision?

To specify the hyperparameters for the training algorithm

To represent different versions of the trained model

To store the images used for training

To define the number of epochs used during training

A custom vision model for detecting defects has precision of 0.80 and recall of 0.90. The cost of a false positive (rejecting a good part) is high, while the cost of a false negative (shipping a defective part) is moderate. The team wants to reduce false positives. What should they do?

Switch to a different model type

Raise the classification threshold

Lower the classification threshold

Add more defect images to the training set

Microsoft Certified:Azure AI Engineer Associate

Domain 4Objective 2

Implement Custom Vision Models AI-102 Practice Questions (Page 4)

Part of the Implement computer vision solutions domain, which accounts for 10-15% of the AI-102 exam. Microsoft does not publish an official question count, but from its 100-minute exam (~40–65 total, ~4–10 in this domain), expect 1–3 from this objective — we provide 36 practice questions to prepare you well beyond it. (estimate)

36questions here

8free pages

7concepts

10-15%of the exam

Questions 16–20

16foundation · easy
A retail company wants to build a custom vision model that identifies whether a product image contains a damaged item or a non-damaged item. The model only needs to output a single label for the entire image. Which type of custom vision model should be used?
Select an answer first
17foundation · easy
A data scientist is labeling images for a custom image classification model that will distinguish between cats and dogs. What is the correct labeling approach?
Select an answer first
18expert · hard
An application is calling a published custom vision prediction endpoint. The response includes multiple tags with probabilities. The application needs to decide whether to act on a prediction. What should it do?
Select an answer first
19foundation · easy
What is the purpose of the 'Training iterations' in Azure Custom Vision?
Select an answer first
20expert · hard
A custom vision model for detecting defects has precision of 0.80 and recall of 0.90. The cost of a false positive (rejecting a good part) is high, while the cost of a false negative (shipping a defective part) is moderate. The team wants to reduce false positives. What should they do?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to AI-102

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by Microsoft. “AI-102” is a trademark of its owner, used for identification only.