Free Microsoft Azure AI Fundamentals Questions and Answers — Questions and Answers

Question 1: Which two tasks are carried out during an Azure Machine Learning process's data ingestion and data preparation phases?

Score test data by using the model.
Combine multiple datasets. (Correct answer)
Remove records that have missing values. (Correct answer)
Calculate the accuracy of the model.

Correct answer: Combine multiple datasets.

During the data ingestion and preparation phases of an Azure Machine Learning process, the focus is on collecting and cleaning data. Combining multiple datasets (data integration) is essential for creating a comprehensive dataset, and removing records with missing values (data cleaning) addresses data quality issues. These steps ensure the data is suitable and robust for subsequent model training.

Question 2: An area's animal population needs to be predicted. What kind of Azure Machine Learning should you employ?

classification
regression (Correct answer)
clustering
None of these

Correct answer: regression

Predicting an area's animal population involves forecasting a continuous numerical value (the count of animals). Regression models are specifically designed for tasks where the output is a numerical quantity. Classification is used for predicting discrete categories, and clustering is for grouping data, neither of which fits the requirement of predicting a specific population count.

Question 3: Which two programming languages are supported by the Azure Machine Learning designer?

C#
Scala
Python (Correct answer)
R (Correct answer)

Correct answer: Python

The Azure Machine Learning designer, while primarily a visual interface, supports custom code execution to extend its capabilities. Python and R are the two most widely adopted programming languages in data science and machine learning. Azure ML provides integrated environments and modules for users to incorporate scripts written in both Python and R into their machine learning pipelines.

Question 4: Your business plans to construct a bottle recycling machine. Bottles with the proper shape must be automatically recognized by the recycling system, while all other objects must be rejected. What kind of AI workload should the business employ?

conversational AI
computer vision (Correct answer)
natural language processing
anomaly detection

Correct answer: computer vision

The task of automatically recognizing bottles of a proper shape and rejecting other objects is a classic application of computer vision. Computer vision AI workloads enable systems to 'see,' process, and interpret visual information from images or videos. This capability is crucial for tasks like object recognition, classification, and shape analysis in real-world scenarios.

Question 5: You have a dataset that details taxi trips that took place over a specific time period. In order to estimate the cost of a taxi ride, a model needs to be trained. What kind of feature should you use?

the fare of individual taxi journeys
the fare of individual taxi journeys
the number of taxi journeys in the dataset
the trip distance of individual taxi journeys (Correct answer)

Correct answer: the trip distance of individual taxi journeys

To accurately estimate the cost (fare) of a taxi ride, the model requires features that are highly correlated with the fare. Trip distance is a primary determinant of taxi fare, as longer distances typically incur higher costs. Therefore, the trip distance of individual taxi journeys is a crucial input feature for training such a predictive model.

Question 6: Which two situations are appropriate for using the Form Recognizer service?

Identify the retailer from a receipt. (Correct answer)
Translate a form from French to English.
Extract the invoice number from an invoice. (Correct answer)
Find image of product in a catalog.

Correct answer: Identify the retailer from a receipt.

Azure Form Recognizer is an AI service specifically designed to extract key-value pairs, text, and tables from various documents. Identifying a retailer from a receipt and extracting an invoice number from an invoice are prime examples of its capabilities, as it specializes in understanding the structure and content of such business documents to automate data entry and processing.

Question 7: You must create a smartphone application so that employees may scan and store their trip expenses. Which kind of computer vision ought to be employed?

image classification
optical character recognition (OCR) (Correct answer)
semantic segmentation
object detection

Correct answer: optical character recognition (OCR)

Scanning and storing trip expenses from physical documents like receipts involves converting images of text into machine-readable text. Optical Character Recognition (OCR) is the specific computer vision technology designed for this purpose. OCR allows the application to extract textual information from the scanned images, making it searchable and usable for expense tracking.

Question 8: A business hires a group of customer care representatives to assist clients by phone and email. The business creates a webchat bot to offer pre-written responses to frequent client questions. What commercial advantages may the corporation anticipate from developing the webchat bot solution?

a reduced workload for the customer service agents (Correct answer)
improved product reliability
increased sales
All of the above

Correct answer: a reduced workload for the customer service agents

A webchat bot designed to answer frequent client questions can automate responses to common inquiries. This automation significantly reduces the number of routine tasks that human customer service agents need to handle. Consequently, it leads to a reduced workload for the agents, allowing them to focus on more complex or unique customer issues.

Question 9: How should data be divided for machine learning training and evaluating?

Use labels for training and features for evaluation.
Randomly split the data into columns for training and columns for evaluation.
Use features for training and labels for evaluation.
Randomly split the data into rows for training and rows for evaluation. (Correct answer)

Correct answer: Randomly split the data into rows for training and rows for evaluation.

For robust machine learning model training and evaluation, it is standard practice to split the dataset into distinct training and testing (or evaluation) sets. This split should be done randomly across the rows (individual data points or observations) to ensure both sets are representative of the overall data distribution and prevent data leakage, allowing the model to be tested on unseen data.

Question 10: Using the automated machine learning user interface, you create a machine learning model (UI). Make sure the model adheres to Microsoft's transparency guideline for ethical AI. What ought you to do?

Enable Explain best model. (Correct answer)
Set Validation type to Auto.
Set Max concurrent iterations to 0.
Set Primary metric to accuracy.

Correct answer: Enable Explain best model.

Microsoft's transparency guideline for ethical AI emphasizes understanding how an AI model makes its decisions. Enabling 'Explain best model' in Azure Machine Learning's automated ML UI generates explanations for the model's predictions, highlighting the features that contributed most to an outcome. This directly addresses the transparency principle by making the model's behavior more interpretable and understandable.

Question 11: You are creating an AI system that empowers everyone, even those who are visually impaired, hard of hearing, or have other disabilities. Which Microsoft AI guiding principle is exemplified by this situation?

inclusiveness (Correct answer)
accountability
fairness
reliability and safety

Correct answer: inclusiveness

Creating an AI system that empowers everyone, including those with visual impairments, hearing impairments, or other disabilities, directly aligns with the Microsoft AI guiding principle of inclusiveness. This principle focuses on designing AI systems that are accessible and beneficial to the widest possible range of people, regardless of their abilities or background, ensuring equitable access and participation.

Question 12: You're constructing an AI system. Which job needs to be added to the service to guarantee that it complies with the Microsoft transparency principle for responsible AI?

Provide documentation to help developers debug code. (Correct answer)
Ensure that a training dataset is representative of the population.
Enable autoscaling to ensure that a service scales based on demand.
Ensure that all visuals have an associated text that can be read by a screen reader.

Correct answer: Provide documentation to help developers debug code.

The transparency principle in responsible AI emphasizes that AI systems should be understandable and their decision-making processes explainable. Providing comprehensive documentation to help developers debug code contributes to transparency by making the system's internal workings and potential issues clearer. This allows for better understanding, auditing, and improvement of the AI system, fostering trust and accountability.

Question 13: Your business is looking into using speech recognition technology into its smart home products. The business wants to find any obstacles that can accidentally exclude certain user groups. Which Microsoft tenet for responsible AI does this represent?

privacy and security
inclusiveness (Correct answer)
fairness
accountability

Correct answer: inclusiveness

Investigating obstacles that might accidentally exclude certain user groups when integrating speech recognition technology directly addresses the principle of inclusiveness. This Microsoft tenet for responsible AI ensures that AI systems are designed to be accessible and fair to all users, preventing bias or limitations that might disadvantage specific demographics or individuals, thereby promoting equitable access and benefit.

Question 14: The criteria used to make the judgment should be comprehensible when designing an AI system to determine if loans should be accepted. Which Microsoft AI guiding principle is exemplified by this situation?

privacy and security
fairness
inclusiveness
transparency (Correct answer)

Correct answer: transparency

The requirement that the criteria used to make judgments, such as loan acceptance, should be comprehensible directly embodies the Microsoft AI guiding principle of transparency. Transparency means that AI systems should be understandable, and their decision-making processes should be explainable to users. This fosters trust and allows for accountability by making the system's logic clear.

Question 15: You are developing an app with AI. Make sure the app adheres to the responsible AI guidelines. Which two guidelines ought to you abide by?

Implement a process of AI model validation as part of the software review process (Correct answer)
Prevent the disclosure of the use of AI-based algorithms for automated decision making
Implement an Agile software development methodology
Establish a risk governance committee that includes members of the legal team, members of the risk management team, and a privacy officer (Correct answer)

Correct answer: Implement a process of AI model validation as part of the software review process

Adhering to responsible AI guidelines involves both technical and organizational measures. Implementing a robust AI model validation process ensures the model's performance, fairness, and reliability are continuously checked. Additionally, establishing a risk governance committee with diverse expertise (legal, risk management, privacy) provides essential oversight and ensures ethical considerations are addressed throughout the AI system's lifecycle, promoting accountability and responsible development.

Question 16: Which kind of machine learning should you employ to forecast the quantity of gift cards that will be sold over the course of the upcoming month?

clustering
regression (Correct answer)
classification
None of these

Correct answer: regression

Forecasting the quantity of gift cards sold over the upcoming month involves predicting a continuous numerical value. Regression models are specifically designed for tasks where the output is a numerical quantity. Classification is used for predicting discrete categories, and clustering is for grouping data, neither of which fits the requirement of predicting a specific numerical quantity.

Free Microsoft Azure AI Fundamentals Questions and Answers

Which two tasks are carried out during an Azure Machine Learning process's data ingestion and data preparation phases?