AI-900 Certification Cheat Sheet

Podcast

Play an AI-generated podcast conversation about this lesson

Questions and Answers

Which of the following is NOT a core principle of Microsoft's Responsible AI Commitment?

Accountability

Equity

Transparency

Profitability (correct)

Which of the following is NOT considered a benefit of MLOps?

Improved security for sensitive data (correct)

Fostered cross-functional collaboration

Enhanced workflow management

Reduced manual efforts in model deployment

Which of the following is NOT a step involved in the Responsible AI Implementation Framework?

Measure Harm Occurrences

Identify Potential Harms

Mitigate Identified Harms

Develop Data Governance Policies (correct)

What is the primary purpose of model interpretability in the context of Responsible AI?

To understand the algorithmic decisions made by the model Signup and view all the answers

What is the role of stakeholders in identifying potential harms?

To provide feedback and insights to inform the assessment process Signup and view all the answers

Which of the following is NOT a level at which mitigations can be implemented to address identified harms?

User level Signup and view all the answers

Why is it important to continuously monitor and update measurement results for harm occurrences?

to ensure the effectiveness and adaptability of mitigations Signup and view all the answers

What is the main objective of establishing mechanisms for user feedback and incident response?

To identify and address potential harms caused by the AI system Signup and view all the answers

What is a key feature of Azure's automated machine learning capabilities?

Automated model selection and tuning Signup and view all the answers

Which Azure service is primarily focused on computer vision tasks involving the detection of objects?

Azure AI Vision Signup and view all the answers

What is a function of Optical Character Recognition (OCR) in Azure AI services?

Converting images of text into machine-encoded text Signup and view all the answers

Which of the following is NOT a common type of generative AI solution in Azure?

Static data analysis Signup and view all the answers

Which component is essential for MLOps in Azure?

Model deployment management Signup and view all the answers

What does the Azure AI Face service primarily provide functionalities for?

Face detection and attributes analysis Signup and view all the answers

What is a major aspect of responsible AI practices in Azure OpenAI Service?

Bias detection and mitigation Signup and view all the answers

Which of the following is a feature of generative AI machine learning in Azure?

Generative design of artifacts Signup and view all the answers

Which application of NLP helps users interact by understanding verbal instructions?

Voice-Activated Assistants Signup and view all the answers

What is the primary challenge of NLP due to the varying ways individuals communicate?

Expression Diversity Signup and view all the answers

Which NLP task involves classifying words according to their grammatical roles?

Part-of-Speech Tagging Signup and view all the answers

How does NLP contribute to global connectivity?

By enabling Language Translation Tools Signup and view all the answers

Which is a key benefit of using summarization algorithms in NLP?

They condense lengthy texts for easy comprehension Signup and view all the answers

What does interpretation complexity in NLP primarily relate to?

The challenges posed by human language ambiguities Signup and view all the answers

What role does entity recognition play in NLP?

To identify and classify text entities Signup and view all the answers

Which of the following is NOT typically a function of NLP?

Data encryption Signup and view all the answers

Which method is specifically utilized for classification tasks with associated probabilities?

Logistic Regression Signup and view all the answers

What is the primary goal of unsupervised learning algorithms?

To uncover inherent patterns in data Signup and view all the answers

What distinguishes Random Forest from other algorithms?

It combines multiple decision trees. Signup and view all the answers

Which algorithm is known for its computational efficiency in classification tasks?

Naive Bayes Classifier Signup and view all the answers

Which technique is employed to condense data while maintaining essential information?

Dimensionality Reduction (PCA) Signup and view all the answers

Which model is specifically tailored for sequential data processing tasks?

Recurrent Neural Networks (RNN) Signup and view all the answers

What aspect of ensemble learning algorithms is defined by combining multiple models?

Overall performance enhancement Signup and view all the answers

Which time series forecasting algorithm effectively handles time series data?

Exponential Smoothing Signup and view all the answers

What is the primary function of automated machine learning in Azure ML?

To streamline model selection and hyperparameter tuning Signup and view all the answers

Which of the following functionalities can Azure ML provide for model deployment?

Deploy models in both web services and containers Signup and view all the answers

How does Azure ML enhance workflow integration with other Azure services?

By enabling seamless merging with several Azure services Signup and view all the answers

What type of models can be developed for predictive analytics using Azure ML?

Models for tasks like sales projection and customer churn prediction Signup and view all the answers

What security measures does Azure ML incorporate for data and models?

Robust security protocols to safeguard data and compliance Signup and view all the answers

In supervised learning, what is the primary role of labelled datasets?

To train models to understand relationships between inputs and outputs Signup and view all the answers

Which application is NOT a typical use case for Azure ML?

Weather forecasting Signup and view all the answers

What fundamental purpose do machine learning algorithms serve?

To identify patterns and forecast outcomes from data Signup and view all the answers

What is the primary purpose of labeled datasets in machine learning?

To provide ground truth information for model training Signup and view all the answers

Which of the following is NOT a prerequisite for creating labeled datasets using Azure Machine Learning?

An understanding of deep learning frameworks Signup and view all the answers

After finishing a data labeling project, what format can data labels be exported to for use in Azure ML?

COCO Format Signup and view all the answers

What method is used to load labeled datasets into a pandas DataFrame within Azure Machine Learning?

to_pandas_dataframe() Signup and view all the answers

Which of the following frameworks is commonly used to build sophisticated machine learning models?

TensorFlow Signup and view all the answers

What role do accelerators like ONNX Runtime serve in deep learning frameworks?

They enhance the performance of the frameworks. Signup and view all the answers

What is the first step to create a labeled dataset using Azure Machine Learning?

Install the Azure Machine Learning SDK for Python. Signup and view all the answers

Which of the following is a function of labeled datasets in Azure Machine Learning?

They enable models to learn and make accurate predictions. Signup and view all the answers

Study Notes

AI-900 Cheat Sheet

Information is for educational purposes only and is not a substitute for official documentation.
The cheat sheet is a resource to help prepare for the AI-900 certification.
The information can be reused, reproduced, and printed but appropriate sources must be credited.
Microsoft Azure provides a free test on Whizlabs website which is helpful for self-assessment.
Hands-on labs and a cloud sandbox environment are available.

Index

The index provides a table of contents to the cheat sheet's topics, which cover different domains.
The page numbers denote the location of specific topics, facilitating quick navigation.

Al Workloads:

The Azure Al Service Platform integrates tools and APIs for AI model construction, deployment, and management.
It's designed for developers, data scientists, and business analysts.
Azure Al Vision leverages advanced CV capabilities for image and video analysis (e.g., image tagging, OCR, facial recognition.)
It enhances content discoverability through image analysis and real-time spatial analysis, empowering applications across various domains (e.g., movement monitoring and retail analytics.)

Real-Time Spatial Analysis

Movement monitoring in real-time is crucial for surveillance, retail analytics, and smart spaces.
Real-time tools are important for extraction of text from images.
Facial recognition enables secure user experiences for authentication.

Azure Content Moderator

Use cases include online marketplaces, gaming, social media, and education sectors.
Key features include text analysis (detecting harmful content like hate speech, violence, and self-harm), image moderation, and video content management.

Language Processing (NLP)

NLP transforms communication between machines and humans by using natural languages.
Applications include voice-activated assistants, language translation tools, emotional analysis, chatbots, summarization tools, and entity recognition.
NLP tasks include speech recognition, part-of-speech tagging, word sense disambiguation, and language modeling.

Knowledge Mining

Knowledge mining learns from vast data volumes to gain insights, identify patterns, and discover relationships within information.
Components include semantic search, content ingestion, content enrichment, and exploration and discovery.
Applications include content research and auditing, risk, and compliance management.

Azure Machine Learning

A range of tools and services are tailored for data scientists, developers, and business analysts.
It facilitates complete, end-to-end ML workflows.
Capabilities include model training and experimentation, automated ML, model deployment, and integration with Azure services.
Use cases include predictive analytics, computer vision, and anomaly detection.

Azure Al Document Intelligence

Automate text and structure extraction from documents with AI and OCR technology.
Streamlined extraction convert documents into actionable data for better decision-making.
Key features include automated extraction, and focus optimization.

Responsible Al:

This section focuses on ethical principles and standards in AI development and deployment.
Azure AI places emphasis on responsible AI development and deployment.
The AI system addresses concerns of bias, equity, and privacy.
Various tools and features are in place to prevent harmful outputs and address ethical concerns.

Generative Al

Generative AI is a specialized subset of AI that creates new content based on existing data.
It produces diverse outputs including images, code, text.

Azure Machine Learning Fairness

Ensuring AI models do not discriminate against specific groups, using fairness metrics like accuracy, error rate, precision, recall, and absolute errors.

MLOPS

MLOps manages the machine learning model lifecycle from development to deployment to maintenance.
Best practices cover version control, continuous integration and delivery (CI/CD), monitoring, and governance and compliance.

Azure Al Vision

Computer vision capabilities for scenarios like identification, contactless access control, and privacy-related face blurring.
Features include image analysis, facial recognition, video content analysis, and OCR.
It offers both custom and pre-built models for diverse applications.

Azure Al Language

Focuses on natural language processing (NLP)
Has various features, including named entity recognition, sentiment analysis, and key phrase extraction.
The platform is accessible through a web interface, REST APIs and Client Libraries, or container deployments.

Azure Al Speech

Azure AI Speech facilitates speech-to-text, text-to-speech, speech translation, and speaker recognition.
Offers advanced customization options and deployment flexibility.

Studying That Suits You

Use AI to generate personalized quizzes and flashcards to suit your learning preferences.

Description

This cheat sheet serves as a study guide for the AI-900 certification exam, detailing the essential topics and resources for mastering Azure AI services. It includes hands-on labs and information on advanced capabilities such as image and video analysis. Use this resource effectively for your certification preparation.