Microsoft Azure AI Fundamentals Quiz

Microsoft Azure AI Fundamentals

• A company employs a team of customer service agents to provide telephone and email support to customers.
  • Creating a webchat bot to provide automated answers to common customer queries can lead to a reduced workload for the customer service agents.

Machine Learning

• For a machine learning process, it's recommended to split data into rows for training and rows for evaluation.
• The Split Data module is useful for separating data into training and testing sets.
  • You can specify the percentage of data to put in each split, but by default, the data is divided 50-50.
  • You can also randomize the selection of rows in each group, and use stratified sampling.

Classification

• In a confusion matrix, True Positive (TP) refers to the class labels in the training set that can take on only two possible values, which are positive or negative.
• False Negative (FN) refers to the incorrectly classified instances.

Model Explainability

• The automated machine learning user interface (UI) enables model explainability, which is critical for building trust and transparency in AI systems.
• Model explainability helps to understand the relationship between input variables (features) and model output.
• It's essential to comply with regulations and best practices, especially in heavily regulated industries like healthcare and banking.

Anomaly Detection

• Anomaly detection encompasses various important tasks in machine learning, such as:
  • Identifying transactions that are potentially fraudulent.
  • Learning patterns that indicate a network intrusion has occurred.
  • Finding abnormal clusters of patients.
  • Checking values entered into a system.

Natural Language Processing (NLP)

• NLP is used for tasks such as:
  • Sentiment analysis
  • Topic detection
  • Language detection
  • Key phrase extraction
  • Document categorization
• NLP enables software to:
  • Analyze text documents to extract key phrases and recognize entities.
  • Perform sentiment analysis to determine the positivity or negativity of language used.
  • Interpret spoken language and synthesize speech responses.
  • Automatically translate spoken or written phrases between languages.

Microsoft Guiding Principles for Responsible AI

• The principles include:
  • Fairness
  • Inclusiveness
  • Transparency
  • Privacy and Security
  • Reliability and Safety
  • Accountability
• Inclusiveness means that AI systems should be designed to incorporate and address a broad range of human needs and experiences.
• Reliability and Safety ensure that AI systems operate reliably, safely, and consistently under normal circumstances and in unexpected conditions.

Feature Engineering

• Feature engineering is applied to generate additional features, and then feature selection is done to eliminate irrelevant, redundant, or highly correlated features.### Azure Machine Learning
• Azure Machine Learning is a cloud-based platform for building, training, and deploying machine learning models.
• It provides a range of tools and services for data scientists, developers, and business users to build, train, and deploy machine learning models.

Trusted AI

• Trusted AI is a set of principles and guidelines for building and deploying AI models that are responsible, reliable, and secure.
• It includes principles such as fairness, transparency, accountability, and explainability.

Responsible AI Principles

• There are four guiding principles for responsible AI: fairness, reliability, safety, and transparency.
• These principles are essential for building and deploying AI models that are trustworthy and responsible.

Machine Learning Tasks

• Machine learning tasks can be classified into three categories: classification, regression, and clustering.
• Classification involves predicting a categorical label or class.
• Regression involves predicting a continuous value or range.
• Clustering involves grouping similar data points into clusters.

Model Evaluation

• Model evaluation is an essential step in the machine learning workflow.
• It involves evaluating the performance of a model on a test dataset.
• Evaluation metrics include accuracy, precision, recall, F1 score, and mean squared error.

Azure Machine Learning Designer

• Azure Machine Learning Designer is a visual interface for building and deploying machine learning models.
• It allows users to create, train, and deploy models using a drag-and-drop interface.
• Users can connect datasets and modules to create a pipeline draft.

Inference Pipeline

• An inference pipeline is a pipeline that is used to make predictions on new data.
• It can be deployed as a real-time endpoint or a batch endpoint.
• Real-time endpoints are used for real-time inferencing, while batch endpoints are used for batch processing.

Automated Machine Learning

• Automated machine learning (AutoML) is a process that automates the machine learning workflow.
• It allows users to automate the selection of algorithms, hyperparameter tuning, and model training.
• AutoML is a no-code solution that can be used to build and deploy machine learning models.

Data Labeling

• Data labeling is the process of annotating data to prepare it for machine learning model training.
• It involves adding labels or annotations to the data to indicate the target value or class.
• Data labeling is an essential step in the machine learning workflow.

Form Recognizer

• Form Recognizer is a service that uses machine learning to extract text, key-value pairs, and tables from documents.
• It can be used to automate the extraction of information from documents.
• Form Recognizer is a part of the Azure Cognitive Services suite.

Prediction

• Prediction is the process of using a machine learning model to make predictions on new data.
• It involves using the trained model to make predictions on new, unseen data.
• Prediction is an essential step in the machine learning workflow.

Azure Kubernetes Service (AKS)

• AKS is a managed container orchestration service that allows users to deploy and manage containerized applications.
• It is used to deploy real-time endpoints for machine learning models.
• AKS provides a scalable and secure environment for deploying machine learning models.

Azure Container Service

• Azure Container Service is a managed container orchestration service that allows users to deploy and manage containerized applications.
• It is used to deploy machine learning models as a service.
• Azure Container Service provides a scalable and secure environment for deploying machine learning models.### Machine Learning Concepts
• Classification: A type of machine learning used to predict a label or category from a set of features. (Example: predicting brain haemorrhage types from brain scan images)
• Regression: A type of machine learning used to predict a numeric value from a set of features. (Example: predicting animal population)
• Clustering: A type of machine learning used to group instances of data into clusters with similar characteristics. (Example: identifying groups of people with similar purchasing habits)

Azure Machine Learning

• Automated ML: A type of machine learning that automatically selects and trains models for a given dataset.
• Designer: A visual interface for building machine learning models without writing code.
• Workspace: A container for machine learning projects, datasets, and models. (Basic and Enterprise workspaces are available)

Data Preparation and Ingestion

• Data Split: Divide a dataset into training and testing sets.
• Feature Engineering: The process of selecting and transforming raw data into features that are suitable for modeling.
• Data Ingestion: The process of loading data into a machine learning environment.

Model Evaluation

• Metrics: Used to evaluate the performance of a machine learning model. (Example: R2, RMSE, F1 score, AUC)
• Cross-Validation: A technique used to evaluate a model by splitting the data into multiple subsets and training/testing the model on each subset.

Azure Cognitive Services

• Custom Vision: A service that allows users to build, deploy, and improve custom image recognition models.
• Computer Vision: A service that provides pre-trained models for image analysis and processing.

Model Training and Deployment

• Training Dataset: A dataset used to train a machine learning model.
• Validation Dataset: A dataset used to evaluate the performance of a model during training.
• Testing Dataset: A dataset used to evaluate the final performance of a trained model.
• Deploying a Model: The process of deploying a trained model into a production environment.

Microsoft Azure AI Fundamentals

• A company employs a team of customer service agents to provide telephone and email support to customers.
  • Creating a webchat bot to provide automated answers to common customer queries can lead to a reduced workload for the customer service agents.

Machine Learning

• For a machine learning process, it's recommended to split data into rows for training and rows for evaluation.
• The Split Data module is useful for separating data into training and testing sets.
  • You can specify the percentage of data to put in each split, but by default, the data is divided 50-50.
  • You can also randomize the selection of rows in each group, and use stratified sampling.

Classification

• In a confusion matrix, True Positive (TP) refers to the class labels in the training set that can take on only two possible values, which are positive or negative.
• False Negative (FN) refers to the incorrectly classified instances.

Model Explainability

• The automated machine learning user interface (UI) enables model explainability, which is critical for building trust and transparency in AI systems.
• Model explainability helps to understand the relationship between input variables (features) and model output.
• It's essential to comply with regulations and best practices, especially in heavily regulated industries like healthcare and banking.

Anomaly Detection

• Anomaly detection encompasses various important tasks in machine learning, such as:
  • Identifying transactions that are potentially fraudulent.
  • Learning patterns that indicate a network intrusion has occurred.
  • Finding abnormal clusters of patients.
  • Checking values entered into a system.

Natural Language Processing (NLP)

• NLP is used for tasks such as:
  • Sentiment analysis
  • Topic detection
  • Language detection
  • Key phrase extraction
  • Document categorization
• NLP enables software to:
  • Analyze text documents to extract key phrases and recognize entities.
  • Perform sentiment analysis to determine the positivity or negativity of language used.
  • Interpret spoken language and synthesize speech responses.
  • Automatically translate spoken or written phrases between languages.

Microsoft Guiding Principles for Responsible AI

• The principles include:
  • Fairness
  • Inclusiveness
  • Transparency
  • Privacy and Security
  • Reliability and Safety
  • Accountability
• Inclusiveness means that AI systems should be designed to incorporate and address a broad range of human needs and experiences.
• Reliability and Safety ensure that AI systems operate reliably, safely, and consistently under normal circumstances and in unexpected conditions.

Feature Engineering

• Feature engineering is applied to generate additional features, and then feature selection is done to eliminate irrelevant, redundant, or highly correlated features.### Azure Machine Learning
• Azure Machine Learning is a cloud-based platform for building, training, and deploying machine learning models.
• It provides a range of tools and services for data scientists, developers, and business users to build, train, and deploy machine learning models.

Trusted AI

• Trusted AI is a set of principles and guidelines for building and deploying AI models that are responsible, reliable, and secure.
• It includes principles such as fairness, transparency, accountability, and explainability.

Responsible AI Principles

• There are four guiding principles for responsible AI: fairness, reliability, safety, and transparency.
• These principles are essential for building and deploying AI models that are trustworthy and responsible.

Machine Learning Tasks

• Machine learning tasks can be classified into three categories: classification, regression, and clustering.
• Classification involves predicting a categorical label or class.
• Regression involves predicting a continuous value or range.
• Clustering involves grouping similar data points into clusters.

Model Evaluation

• Model evaluation is an essential step in the machine learning workflow.
• It involves evaluating the performance of a model on a test dataset.
• Evaluation metrics include accuracy, precision, recall, F1 score, and mean squared error.

Azure Machine Learning Designer

• Azure Machine Learning Designer is a visual interface for building and deploying machine learning models.
• It allows users to create, train, and deploy models using a drag-and-drop interface.
• Users can connect datasets and modules to create a pipeline draft.

Inference Pipeline

• An inference pipeline is a pipeline that is used to make predictions on new data.
• It can be deployed as a real-time endpoint or a batch endpoint.
• Real-time endpoints are used for real-time inferencing, while batch endpoints are used for batch processing.

Automated Machine Learning

• Automated machine learning (AutoML) is a process that automates the machine learning workflow.
• It allows users to automate the selection of algorithms, hyperparameter tuning, and model training.
• AutoML is a no-code solution that can be used to build and deploy machine learning models.

Data Labeling

• Data labeling is the process of annotating data to prepare it for machine learning model training.
• It involves adding labels or annotations to the data to indicate the target value or class.
• Data labeling is an essential step in the machine learning workflow.

Form Recognizer

• Form Recognizer is a service that uses machine learning to extract text, key-value pairs, and tables from documents.
• It can be used to automate the extraction of information from documents.
• Form Recognizer is a part of the Azure Cognitive Services suite.

Prediction

• Prediction is the process of using a machine learning model to make predictions on new data.
• It involves using the trained model to make predictions on new, unseen data.
• Prediction is an essential step in the machine learning workflow.

Azure Kubernetes Service (AKS)

• AKS is a managed container orchestration service that allows users to deploy and manage containerized applications.
• It is used to deploy real-time endpoints for machine learning models.
• AKS provides a scalable and secure environment for deploying machine learning models.

Azure Container Service

• Azure Container Service is a managed container orchestration service that allows users to deploy and manage containerized applications.
• It is used to deploy machine learning models as a service.
• Azure Container Service provides a scalable and secure environment for deploying machine learning models.### Machine Learning Concepts
• Classification: A type of machine learning used to predict a label or category from a set of features. (Example: predicting brain haemorrhage types from brain scan images)
• Regression: A type of machine learning used to predict a numeric value from a set of features. (Example: predicting animal population)
• Clustering: A type of machine learning used to group instances of data into clusters with similar characteristics. (Example: identifying groups of people with similar purchasing habits)

Azure Machine Learning

• Automated ML: A type of machine learning that automatically selects and trains models for a given dataset.
• Designer: A visual interface for building machine learning models without writing code.
• Workspace: A container for machine learning projects, datasets, and models. (Basic and Enterprise workspaces are available)

Data Preparation and Ingestion

• Data Split: Divide a dataset into training and testing sets.
• Feature Engineering: The process of selecting and transforming raw data into features that are suitable for modeling.
• Data Ingestion: The process of loading data into a machine learning environment.

Model Evaluation

• Metrics: Used to evaluate the performance of a machine learning model. (Example: R2, RMSE, F1 score, AUC)
• Cross-Validation: A technique used to evaluate a model by splitting the data into multiple subsets and training/testing the model on each subset.

Azure Cognitive Services

• Custom Vision: A service that allows users to build, deploy, and improve custom image recognition models.
• Computer Vision: A service that provides pre-trained models for image analysis and processing.

Model Training and Deployment

• Training Dataset: A dataset used to train a machine learning model.
• Validation Dataset: A dataset used to evaluate the performance of a model during training.
• Testing Dataset: A dataset used to evaluate the final performance of a trained model.
• Deploying a Model: The process of deploying a trained model into a production environment.