AI-900 Practice Test 48 — 15 Questions

Question 1

A data scientist is developing a classification model to detect fraudulent transactions. The dataset is split into training and test sets. The data scientist repeatedly tunes the model's hyperparameters and evaluates performance on the test set until the test accuracy reaches 95%. However, when the model is deployed on new, unseen data, its accuracy drops to 70%. Which concept best explains this performance degradation?

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

Accepted Answer

Overfitting to the test set. The data scientist repeatedly tuned hyperparameters based on test set performance, effectively using the test set as part of the training process. This causes the model to become specialized to the test set's specific patterns and noise, so it fails to generalize to new, unseen data. This phenomenon is known as overfitting to the test set, where the test set no longer provides an unbiased estimate of real-world performance.

Answer

Overfitting to the training data

Answer

Data leakage from the training set to the test set

Answer

Underfitting the training data

Question 2

A hospital uses an AI system to recommend patient treatment plans. A doctor questions why the system recommended a specific treatment for a particular patient. Which Microsoft responsible AI principle is most directly relevant to providing the answer?

Accepted Answer

Transparency. Transparency is the responsible AI principle that requires AI systems to be understandable and interpretable by humans. When a doctor questions why a specific treatment was recommended, the system must be able to provide an explanation of its reasoning, such as which patient features (e.g., lab results, medical history) most influenced the recommendation. This aligns with the need for explainability in AI, enabling clinicians to trust and validate the model's output.

Answer

Fairness

Answer

Reliability and Safety

Answer

Privacy and Security

Question 3

What is the Azure AI Custom Vision portal used for?

Accepted Answer

Training and evaluating custom image classification and object detection models without code. The Azure AI Custom Vision portal is a no-code web interface that allows users to upload images, label them, and train custom image classification or object detection models. It abstracts away the underlying machine learning code, making it accessible for non-developers to build and evaluate computer vision models tailored to their specific use cases.

Answer

Managing Azure subscription billing for AI services

Answer

Building chatbots using natural language understanding

Answer

Monitoring the health of deployed AI services

Question 4

What is 'credit scoring' as an AI workload and what responsible AI concerns does it raise?

Accepted Answer

ML for predicting loan repayment risk — with fairness, bias, and explainability concerns. Credit scoring in AI refers to machine learning models that predict the likelihood of a borrower repaying a loan. This raises responsible AI concerns around fairness (e.g., models may discriminate against protected groups), bias (e.g., training data may reflect historical inequalities), and explainability (e.g., complex models like gradient-boosted trees are often black boxes, making it hard to justify decisions to regulators or customers).

Answer

A system for automatically assigning credit scores to software bugs in a development backlog

Answer

Monitoring whether a customer has used all their credit within an approved limit

Answer

An internal system for scoring the quality of AI projects within an organisation

Question 5

What is 'zero-shot object detection' in computer vision?

Accepted Answer

Detecting objects described in text without any training examples of that specific class. Zero-shot object detection refers to a model's ability to detect objects in images based on a textual description of the target class, without having been trained on any labeled examples of that specific class. This is achieved by leveraging a joint embedding space where visual features and text features are aligned, allowing the model to generalize to unseen categories at inference time.

Answer

Object detection that runs with zero latency for real-time applications

Answer

Detection that works on black and white images (zero colour channels)

Answer

An object detection model with zero false positives on the test set

Question 6

What is the role of intents in conversational language understanding (CLU)?

Accepted Answer

Intents represent the user's goal or desired action, determining how the bot should respond. In conversational language understanding (CLU), intents represent the user's goal or desired action, such as booking a flight or checking the weather. They map user utterances to specific tasks the bot should perform, enabling the model to classify input and trigger appropriate responses. This is distinct from entities (which extract data) or responses (which are outputs).

Answer

Intents are the specific pieces of information extracted from user messages (dates, amounts, names)

Answer

Intents are the predefined bot responses stored in a knowledge base

Answer

Intents represent the confidence level of a bot's understanding

Question 7

What is the role of a validation dataset in machine learning?

Accepted Answer

To tune hyperparameters and monitor performance during training without using test data. The validation dataset is used during model training to tune hyperparameters and monitor performance on unseen data, preventing overfitting without contaminating the test set. This allows iterative adjustments to model architecture or learning rate while keeping the test data reserved for final evaluation.

Answer

To provide the primary examples for training the model's weights

Answer

To provide the final, unbiased assessment of model performance

Answer

To store the model's trained weights for later use

Question 8

A retail warehouse uses a camera system to locate and count boxes on shelves. The system needs to output the exact positions of each box by drawing a rectangular frame around it in the image. Which Azure Computer Vision capability should they use?

Accepted Answer

Object detection. Object detection is the correct capability because it identifies and localizes multiple objects within an image by drawing bounding boxes around each detected instance. In this scenario, the system needs to locate and count individual boxes on shelves, which requires both classification (what is a box) and localization (where each box is), exactly what object detection provides.

Answer

Image classification

Answer

Semantic segmentation

Answer

Optical Character Recognition (OCR)

Question 9

A data scientist trains a model on historical data and achieves high accuracy on both the training set and a held-out test set. However, when the model is deployed in production, it performs poorly on new, unseen data. Which issue is most likely the cause?

Accepted Answer

Data leakage. Data leakage occurs when information from outside the training dataset is inadvertently used to train the model, causing it to learn patterns that do not generalize to new data. In this scenario, the high accuracy on both training and test sets but poor production performance indicates that the test set was contaminated with information from the future or from the target variable, making the model appear accurate during validation but fail in real-world deployment.

Answer

Overfitting

Answer

Underfitting

Answer

Concept drift

Question 10

A research team wants to automatically analyze thousands of online news articles to identify the main topics discussed in each article (e.g., 'technology', 'politics', 'sports'). They need a prebuilt Azure AI Language feature that returns a list of key concepts or topics without any custom training. Which feature should they use?

Accepted Answer

Key phrase extraction. Key phrase extraction is the correct Azure AI Language feature because it automatically identifies and returns a list of the main topics, concepts, or themes discussed in a document without requiring any custom training or labeled data. This prebuilt capability is designed specifically for extracting high-level topics from unstructured text, making it ideal for analyzing thousands of news articles to determine subjects like 'technology', 'politics', or 'sports'.

Answer

Named Entity Recognition (NER)

Answer

Sentiment analysis

Answer

Language detection

Question 11

What is the Azure AI Language 'orchestration workflow' feature used for?

Accepted Answer

Routing user utterances to the appropriate CLU or QnA component based on intent. The orchestration workflow feature in Azure AI Language is designed to connect multiple Conversational Language Understanding (CLU) and Question Answering (QnA) projects into a single endpoint. It uses a top-level orchestrator model to classify the user's intent and then routes the utterance to the appropriate child project (CLU or QnA) for further processing, enabling a unified conversational experience across different knowledge bases.

Answer

Automating data processing pipelines in Azure Data Factory

Answer

Scheduling NLP model training jobs at regular intervals

Answer

Translating utterances between languages before processing

Question 12

What is a 'compute instance' in Azure Machine Learning?

Accepted Answer

A managed cloud workstation for interactive ML development with pre-installed tools. A compute instance in Azure Machine Learning is a fully managed cloud workstation that provides a pre-configured environment with popular ML tools like Jupyter Notebooks, TensorFlow, and PyTorch. It is designed for interactive development, allowing data scientists to train and experiment with models without managing infrastructure.

Answer

A scalable cluster for running distributed training jobs across many nodes

Answer

A virtual machine that automatically scales to run batch predictions

Answer

A serverless execution environment for ML inference requests

Question 13

An insurance company uses an AI system to automatically process and approve or reject claims. The system sometimes rejects valid claims because the uploaded documents are in slightly different formats (e.g., PDF vs. scanned images). The company wants to minimize these errors. Which Microsoft responsible AI principle is most directly relevant to addressing this issue?

Accepted Answer

Reliability and safety. The issue is that the AI system fails to process valid claims due to variations in document formats (PDF vs. scanned images), which is a reliability and safety problem. The system should be robust enough to handle input variations and consistently produce correct outcomes. Microsoft's Reliability and safety principle focuses on ensuring AI systems operate reliably, safely, and consistently under expected conditions, directly addressing the need to minimize such errors.

Answer

Fairness

Answer

Inclusiveness

Answer

Transparency

Question 14

A parking management company uses cameras at the entrance and exit of a lot. They need to automatically read the license plate numbers of each car as it enters and exits. Which Azure Computer Vision capability is specifically designed for this task?

Accepted Answer

Optical Character Recognition (OCR). Optical Character Recognition (OCR) is the Azure Computer Vision capability specifically designed to extract printed or handwritten text from images, including license plate numbers. In this scenario, the cameras capture images of cars entering and exiting, and OCR processes those images to read the alphanumeric characters on the license plates. This is the exact use case for OCR, as it can handle varied fonts, angles, and lighting conditions common in parking lot environments.

Answer

Object detection

Answer

Image classification

Answer

Facial recognition

Question 15

What is the difference between zero-shot, one-shot, and few-shot learning in prompting?

Accepted Answer

Zero-shot uses no examples; few-shot provides multiple examples in the prompt to guide responses. Zero-shot learning involves providing no examples in the prompt, relying solely on the model's pre-trained knowledge to generate a response, while few-shot learning includes multiple examples (typically 2–5) within the prompt to guide the model's output pattern. This distinction is fundamental to prompt engineering in generative AI workloads on Azure, where the number of examples directly influences output consistency and task specificity without retraining the model.

Answer

They refer to how many GPUs are used for model training

Answer

They refer to how many training epochs the model underwent

Answer

Zero-shot is for beginners; few-shot is for experts