AI-900 Practice Test 41 — 15 Questions

Question 1

A company develops an AI system to screen job candidates based on their resumes. The system is trained on historical data. Analysis reveals that the model has an adverse impact against female candidates due to a proxy feature (e.g., 'years of continuous employment') that correlates with gender. The team removes the protected attribute 'gender' from the training data but the biased outcome persists. According to Microsoft's responsible AI principles, which additional step should the team take to address this unfairness?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Use a tool like Fairlearn to detect and mitigate the bias while maintaining model performance.. Microsoft's responsible AI principle of fairness requires not just removing protected attributes but also detecting and mitigating proxy features that cause bias. Fairlearn is a Microsoft open-source toolkit specifically designed to assess and mitigate unfairness in AI systems, offering algorithms like 'Exponentiated Gradient Reduction' or 'Grid Search' to reduce disparity while preserving model performance. Simply removing the proxy feature (A) may not always be feasible if it carries predictive value, and Fairlearn provides a systematic way to balance fairness and accuracy.

Answer

Remove the offending proxy feature 'years of continuous employment' from the training data.

Answer

Train a separate model for each gender group to ensure equal outcomes.

Answer

Collect more training data from underrepresented groups.

Question 2

A retail store wants to analyze customer behavior in front of a specific product display. They need to determine how long each customer stands in front of the display and whether they pick up an item. Which Azure Computer Vision capability should they use?

Accepted Answer

Spatial Analysis. Spatial Analysis is the correct Azure Computer Vision capability because it is specifically designed to analyze people's movement, presence, and interactions within a physical space using video feeds. It can track how long a customer stands in front of a display (dwell time) and detect actions like picking up an item, by processing bounding boxes and skeleton data from cameras.

Answer

Image Classification

Answer

Optical Character Recognition (OCR)

Answer

Object Detection

Question 3

A medical research team trains a model to detect a rare disease from lab results. The disease occurs in only 1% of patients. The model predicts 'no disease' for every patient and achieves 99% accuracy. Which metric best reveals that the model is failing to identify actual disease cases?

Accepted Answer

Recall. Recall (sensitivity) measures the proportion of actual positive cases correctly identified by the model. With a 99% accuracy but zero true positives (since the model always predicts 'no disease'), recall is 0%, which directly reveals the model's failure to detect any actual disease cases. In Azure Machine Learning, recall is a key metric for imbalanced classification tasks, especially when missing a positive case has severe consequences.

Answer

Accuracy

Answer

Precision

Answer

F1 score

Question 4

A manufacturing company uses Azure Computer Vision to analyze assembly line images. They need to identify specific product defects (e.g., scratches, dents) and also read serial numbers printed on the products in various fonts. Which combination of Azure Computer Vision features should they use?

Accepted Answer

Custom Vision (object detection) and OCR. The scenario requires two distinct capabilities: identifying specific defect types (scratches, dents) and reading variable-font serial numbers. Custom Vision's object detection model can be trained on labeled defect images to recognize those specific patterns, while Azure's OCR (part of Computer Vision's Read API) extracts printed text regardless of font. Combining these two features directly addresses both requirements.

Answer

Image Analysis (object detection) and OCR

Answer

Face API and OCR

Answer

Image Analysis (tags) and OCR

Question 5

A legal department needs to automatically extract specific entities from contracts, such as 'Effective Date', 'Governing Law', and 'Payment Terms'. They have 500 manually labeled contract clauses that specify which text spans correspond to each entity. Which Azure AI Language feature should they use to build this custom entity extraction solution?

Accepted Answer

Custom Named Entity Recognition (Custom NER). Custom Named Entity Recognition (Custom NER) is the correct choice because it allows you to train a model on your own labeled data (the 500 manually labeled contract clauses) to extract domain-specific entities like 'Effective Date', 'Governing Law', and 'Payment Terms' that are not covered by prebuilt models. This feature uses a custom trained model to identify and classify text spans according to your defined schema.

Answer

Prebuilt Named Entity Recognition (NER)

Answer

Key phrase extraction

Answer

Custom text classification

Question 6

A customer support team receives hundreds of long product reviews every day. They want to automatically summarize each review into a few key sentences to quickly understand the main points. Which prebuilt Azure AI Language feature should they use?

Accepted Answer

Extractive summarization. Extractive summarization is the correct choice because it is specifically designed to condense long documents into a few key sentences by extracting the most important sentences directly from the original text. This aligns perfectly with the customer support team's goal of automatically summarizing hundreds of product reviews into concise, key points for quick understanding.

Answer

Key phrase extraction

Answer

Sentiment analysis

Answer

Entity recognition

Question 7

An autonomous vehicle company uses an AI system for navigation. During testing, the system performs well in sunny weather but fails in snowy conditions because the training data had very few examples of snowy roads. The company decides to deploy the system anyway, hoping it will learn on the road. Which Microsoft responsible AI principle is most directly violated by this decision?

Accepted Answer

Reliability and Safety. The decision to deploy an AI system that is known to fail in snowy conditions directly violates the Reliability and Safety principle. This principle requires that AI systems operate reliably and safely under all expected conditions, and that potential failures are identified and mitigated before deployment. By hoping the system will 'learn on the road,' the company is exposing users and the public to unacceptable risk, as the system has not been validated for safe operation in snowy environments.

Answer

Fairness

Answer

Privacy and Security

Answer

Inclusiveness

Question 8

A security company wants to monitor a restricted area using camera feeds. The system must detect if a person is present in each video frame and draw a rectangle around each detected person. Which Azure Cognitive Services Computer Vision capability should they use?

Accepted Answer

Object Detection. Object Detection is the correct capability because it identifies and locates objects (including people) within an image by drawing bounding boxes around each detected instance. This directly matches the requirement to detect persons in video frames and draw rectangles around them, which is a core function of the Object Detection API in Azure Cognitive Services Computer Vision.

Answer

Image Analysis (Describe image)

Answer

Optical Character Recognition (OCR)

Answer

Face Detection

Question 9

A company uses a generative AI model to answer customer questions about their products. They observe that the model sometimes produces factually incorrect or fabricated information. To reduce these inaccuracies, they want to provide the model with relevant, up-to-date product documentation as context before generating a response. Which technique is being applied?

Accepted Answer

Grounding. B is correct because grounding is the technique of providing a generative AI model with specific, authoritative source data (such as product documentation) as context before generating a response. This anchors the model's output to verified facts, directly reducing hallucinations and fabricated information by constraining the generation to the provided context.

Answer

Prompt Engineering

Answer

Fine-tuning

Answer

Reinforcement Learning from Human Feedback (RLHF)

Question 10

A marketing team wants to use Azure OpenAI Service to generate product descriptions that consistently match a specific brand voice. They have a small set of example descriptions that demonstrate the desired tone. They want to adapt the model without retraining it from scratch. Which approach should they take?

Accepted Answer

Use prompt engineering with few-shot learning by including the example descriptions in the prompt. Prompt engineering with few-shot learning allows the model to infer the desired brand voice from the example descriptions included directly in the prompt, without requiring retraining. This approach leverages the model's in-context learning capability, where it adapts its output based on the provided examples while keeping the base model unchanged.

Answer

Fine-tune the base model on the example descriptions

Answer

Increase the temperature parameter to the maximum value

Answer

Train a new model using Azure Machine Learning

Question 11

A retail company wants to automatically group its customers into distinct segments based on their purchasing patterns, without having pre-defined categories. The goal is to discover natural groupings in the customer data to tailor marketing campaigns. Which type of machine learning task should the company use?

Accepted Answer

Unsupervised learning - Clustering. The company wants to discover natural groupings in customer data without pre-defined categories, which is the definition of unsupervised learning. Clustering algorithms (e.g., K-Means, DBSCAN) automatically partition data into segments based on similarity in purchasing patterns, making it the correct choice for this scenario.

Answer

Supervised learning - Classification

Answer

Reinforcement learning

Answer

Supervised learning - Regression

Question 12

A company wants to use Azure OpenAI to generate personalized marketing emails. They have a large dataset of customer purchase histories. They want the model to generate emails that recommend products based on individual customer preferences without retraining the entire model. Which technique should they use?

Accepted Answer

Prompt engineering with few-shot learning. Prompt engineering with few-shot learning is correct because it allows the model to generate personalized marketing emails by providing a few examples of customer-product pairs in the prompt, without modifying the underlying model weights. This technique leverages the pre-trained knowledge of Azure OpenAI to recommend products based on individual customer purchase histories, avoiding the need for costly retraining.

Answer

Fine-tuning

Answer

Reinforcement learning from human feedback

Answer

Creating a custom neural network

Question 13

A marketing team uses Azure OpenAI Service to generate social media posts. They want the generated text to be more creative and diverse, with unexpected word choices. Which parameter should they increase?

Accepted Answer

temperature. Increasing the temperature parameter makes the model more creative and diverse by raising the probability of sampling lower-probability tokens, leading to unexpected word choices. Temperature controls the randomness of token selection, with higher values (e.g., 0.9) producing more varied outputs, which aligns with the team's goal of generating creative social media posts.

Answer

frequency_penalty

Answer

presence_penalty

Answer

top_p

Question 14

A data scientist trains a classification model to predict whether an email is 'phishing' or 'legitimate'. The model achieves 99% accuracy on the training data but only 68% accuracy on the test data. Which action is most likely to help improve the model's generalization performance?

Accepted Answer

Apply regularization techniques such as L1 or L2 regularization.. The model's high training accuracy (99%) paired with much lower test accuracy (68%) is a classic sign of overfitting, where the model has memorized the training data rather than learning generalizable patterns. Regularization techniques like L1 (Lasso) or L2 (Ridge) add a penalty to the loss function that discourages overly complex models by shrinking the weights of less important features, directly reducing overfitting and improving generalization on unseen data.

Answer

Increase the number of training epochs significantly.

Answer

Remove some of the training data to make the dataset smaller.

Answer

Add more layers and neurons to the neural network.

Question 15

A writer uses Azure OpenAI Service to generate multiple story ideas. They find that the model often repeats the same concepts across different outputs. Which parameter should they increase to reduce repetition and encourage more novel content?

Accepted Answer

Frequency penalty. The frequency penalty parameter in Azure OpenAI Service reduces the likelihood of repeating the same tokens or phrases by applying a penalty proportional to the frequency of tokens already generated. Increasing this value discourages the model from reusing common concepts, thereby promoting more novel and diverse story ideas.

Answer

Temperature

Answer

Top_p (nucleus sampling)

Answer

Max tokens