Last updated on June 24, 2024
Here are 10 AI-900 Microsoft Azure AI Fundamentals practice exam questions to help you gauge your readiness for the actual exam.
Question 1
Match the relevant Microsoft guiding principles for responsible AI with their appropriate descriptions.
Instructions: To answer, drag the appropriate item from the column on the left to its description on the right. Each correct match is worth one point.
Correct Answer:
Accountability: Establishing procedures to enable human intervention and override the decisions made by AI systems.
Transparency: Enables the team to comprehend the data, algorithms, transformations, and final model associated with the training process.
Reliability and safety: Guaranteeing that AI systems adhere to their original design, effectively handle unexpected circumstances, and are resistant to malicious manipulation.
Fairness: Ensure non-discrimination based on gender, race, sexual orientation, or religion, Microsoft offers an AI checklist.
Microsoft emphasizes six key principles for responsible AI: accountability, inclusiveness, reliability and safety, fairness, transparency, and privacy and security.
Accountability principle means putting in place procedures that allow humans to intervene and override the decisions made by AI systems. This ensures that there is a responsible entity overseeing the AI system and taking responsibility for its actions, addressing any biases or errors that may arise.
Transparency emphasizes the importance of providing visibility into the data, algorithms, and transformations used in AI models. By doing so, stakeholders can gain a better understanding of how the AI system operates and can scrutinize its processes.
Reliability and safety principle focuses on ensuring that AI systems consistently perform as intended, respond effectively to unforeseen situations, and are safeguarded against manipulation. This is essential for building trust in AI and preventing unintended harm or malicious exploitation.
Fairness principle highlights the importance of avoiding discrimination and bias in AI systems to ensure equal treatment for all individuals. By providing guidelines and checklists, developers and organizations can assess and address potential biases, promoting fairness in AI decision-making processes.
References:
https://learn.microsoft.com/en-us/azure/cloud-adoption-framework/innovate/best-practices/trusted-ai
https://learn.microsoft.com/en-us/azure/machine-learning/concept-responsible-ai?view=azureml-api-2
Check out this Azure Responsible AI Cheat Sheet:
https://tutorialsdojo.com/azure-responsible-ai/
Question 2
What guiding principle ensures that consumers have control over their data and safeguards their privacy and security in AI systems?
- Transparency
- Reliability and Safety
- Privacy and Security
- Accountability
Question 3
What principle in responsible AI emphasize regarding the ability of humans to intervene and override decisions made by AI systems when required?
- Reliability and safety
- Accountability
- Inclusiveness
- Transparency
Question 4
Which responsible AI principle aligns with the solution of using Snapshots within Azure Machine Learning workspaces to record or retrain training-related assets and metrics, thus promoting transparency?
- Reliability and safety
- Accountability
- Inclusiveness
- Transparency
Question 5
Match the types of AI workloads to the appropriate scenarios.
To answer, drag the appropriate workload type from the column on the left to its scenario on the right.
Correct Answer:
Natural language processing (NLP): Translate English to Filipino
Computer vision: Image search engines
Regression: Property assessment valuation
Classification: Property assessment valuation
Azure Machine Learning offers AutoML capabilities in five domains: classification, regression, forecasting, computer vision, and natural language processing (NLP). With AutoML, users can automate the process of building high-quality machine learning models, leveraging breakthroughs from Microsoft Research and streamlining the ML development experience.
Natural language processing (NLP) – Translating English to Filipino is an example of NLP, as it involves analyzing and generating human language. The scenarios of review insights and automated responses also align with NLP’s aim to analyze and understand text or speech data.
Computer vision – Image search engines leverage computer vision to extract meaningful information from images and videos. The mentioned scenarios of interpreting handwriting and pattern recognition fall within the domain of computer vision.
Regression – Property assessment valuation involves analyzing relationships between variables (such as property characteristics) to predict a numerical outcome (property value). This aligns with the regression category.
Classification – Email spam detection is a common scenario of classification where the AI model classifies emails into spam or legitimate categories based on their features.
References:
https://learn.microsoft.com/en-us/azure/machine-learning/concept-automated-ml?view=azureml-api-2
https://azure.microsoft.com/en-us/products/machine-learning/automatedml/#overview
Check out this Automated Machine Learning (AutoML) in Azure Cheat Sheet:
https://tutorialsdojo.com/automated-machine-learning-automl-in-azure/
Question 6
Which specific workload type within AutoML is typically employed in the development process of chatbots that possess the capability to automatically respond to inquiries pertaining to refunds and exchanges?
- Classification
- Regression
- Natural language processing (NLP)
- Computer vision
Correct Answer: 3
Natural Language Processing (NLP) specifically deals with text or speech data and aims to analyze, understand, and generate human language. It is the most suitable choice for developing chatbots that can automatically process and respond to questions.
Hence, the correct answer is: Natural Language Processing (NLP).
Classification is incorrect because it is not the suitable choice for the complex task of understanding and generating human language required for answering questions about refunds and exchanges. While it can be used for intent classification, it lacks the complexities needed for natural language understanding.
Regression is incorrect because it is not directly applicable to processing and understanding natural language for chatbot responses. Its focus is on predicting numerical values based on independent variables, which does not align with the complexities of language processing.
Computer vision is incorrect because it is not directly applicable to understanding and generating human language. It focuses on analyzing and interpreting visual data from images or videos, which is different from processing and responding to language-based queries.
References:
https://learn.microsoft.com/en-us/azure/machine-learning/concept-automated-ml?view=azureml-api-2
https://learn.microsoft.com/en-us/azure/architecture/data-guide/technology-choices/natural-language-processing
Check out this Automated Machine Learning (AutoML) in Azure Cheat Sheet:
https://tutorialsdojo.com/automated-machine-learning-automl-in-azure/
Question 7
You are tasked with developing a predictive model for a multi-location retail business to generate accurate forecasts for future sales, taking into account the historical sales data from each store and the unique seasonal patterns at each location.
In this scenario, which AutoML workload type would you select to build a predictive model that effectively incorporates both the historical sales data and the seasonal variations?
- Classification
- Regression
- Time-series forecasting
- Computer vision
Correct Answer: 3
Time-series forecasting is specifically designed to handle sequential data and capture the time-dependent patterns present in the sales data. By incorporating seasonal variations, it can generate more accurate forecasts for future sales, making it the most appropriate choice in this scenario.
Hence, the correct answer is: Time-series forecasting.
Classification is incorrect because classification is not suitable for forecasting future sales based on historical data and seasonal patterns. Classification is used to categorize data into predefined classes or categories, but it does not capture time-dependent patterns and trends.
Regression is incorrect because while regression can predict numerical values based on independent variables, it may not effectively capture the time-dependent patterns present in sales data with seasonal variations. Regression models assume a linear relationship between variables and may not account for the specific time-series.
Computer vision is incorrect because computer vision is not directly applicable to forecasting future sales. Computer vision is focused on analyzing and interpreting visual data from images or videos, which is not the primary data type involved in sales forecasting.
References:
https://learn.microsoft.com/en-us/azure/machine-learning/concept-automated-ml?view=azureml-api-2
https://learn.microsoft.com/en-us/analysis-services/data-mining/microsoft-time-series-algorithm?view=asallproducts-allversions
Check out this Automated Machine Learning (AutoML) in Azure Cheat Sheet:
https://tutorialsdojo.com/automated-machine-learning-automl-in-azure/
Question 8
Match the types of AI workloads to the appropriate scenarios.
To answer, drag the appropriate workload type from the column on the left to its scenario on the right.
Question 9
You are assigned to develop an ML solution that can accept natural language input and generate original content. The solution should return appropriate responses in a variety of formats, such as in the form of an image, a code or an easy-to-understand natural language response.
Which of the following should you use to meet the above requirements?
- Generative AI
- Predictive AI
- Time-series forecasting
- Natural language processing (NLP)
Correct Answer: 1
Artificial Intelligence (AI) imitates human behavior by using machine learning to interact with the environment and execute tasks without explicit directions on what to output. Generative AI describes a category of capabilities within AI that create original content. People typically interact with generative AI that has been built into chat applications. One popular example of such an application is Microsoft Copilot, a chatbot companion to browse the web more effectively. Generative AI applications take in natural language input and return appropriate responses in a variety of formats, such as natural language, images, or code.
The Azure OpenAI Service has a key advantage in its smooth integration with other Azure services, such as Azure Machine Learning and Azure Cognitive Services, which helps in developing, deploying and scaling AI applications. This service also provides enterprise-grade security and compliance features to protect sensitive data while leveraging powerful AI capabilities. By offering a managed environment for LLMs, Azure OpenAI Service simplifies the process of building and deploying AI models, making it easier for organizations to utilize generative AI to improve their operations and innovate in their respective fields.
Hence, the correct answer is: Generative AI.
Predictive AI is incorrect because this type of AI uses historical data and other sources of information to predict a certain result and not to generate content.
Time-series forecasting is incorrect because it is primarily used for predicting future values based on historical time-based data, such as sales or stock prices. It is not capable of generating new content.
Natural language processing (NLP) is incorrect because its main focus is on understanding and processing human language, which would be more relevant for tasks involving text analysis, sentiment analysis, or language translation. This is actually just one part of Generative AI and is not enough to meet the requirements in itself.
References:
https://learn.microsoft.com/en-us/training/paths/introduction-generative-ai/
https://learn.microsoft.com/en-us/training/modules/fundamentals-generative-ai/2-what-is-generative-ai
Check out this related Generative AI and Foundation Models Cheat Sheet:
https://tutorialsdojo.com/the-ai-revolution-an-overview-of-generative-ai-and-foundation-models/
Question 10
Which responsible AI principle aligns with the solution of implementing algorithmic techniques such as demographic parity or equalized odds to ensure that an AI system does not discriminate based on gender, race, sexual orientation, or religion?
- Reliability and safety
- Accountability
- Inclusiveness
- Fairness
For more practice questions like these and to further prepare you for the actual AI-900 Microsoft Azure AI Fundamentals exam, we recommend that you take our top-notch AI-900 Microsoft Azure AI Fundamentals Practice Exams, which simulate the real unique question types in the AI-900 exam such as drag and drop, dropdown, and hotspot.
Get 20% Off – Christmas Big Sale on All Practice Exams, Video Courses, and eBooks!
Learn AWS with our PlayCloud Hands-On Labs
Tutorials Dojo Exam Study Guide eBooks
FREE AWS Exam Readiness Digital Courses
FREE AWS, Azure, GCP Practice Test Samplers
Subscribe to our YouTube Channel
Follow Us On Linkedin
Written by: Tutorials Dojo
