AZ-900 Episode 16 | Azure Artificial Intelligence (AI) Services | Machine Learning Studio & Service

Adam Marczak - Azure for Everyone
21 Sept 202008:09

TLDRIn this episode, we delve into the realm of Azure Artificial Intelligence (AI) Services, focusing on Machine Learning Studio and Service. We begin with an overview of AI and Machine Learning (ML), explaining how ML enables software to learn from data and make predictions. Azure Machine Learning is highlighted as the key service for building ML models, offering tools such as notebooks in Python or R, a visual designer for drag-and-drop model building, and Automated ML for algorithm selection and parameter tuning. The episode demonstrates the end-to-end process of model training, validation, deployment, and monitoring within Azure Machine Learning Studio, emphasizing the ease of use and the comprehensive management of compute resources, data stores, and model deployment. The Azure Machine Learning workspace is introduced as the central hub for all ML operations, including asset management and external service connections. The episode concludes with a quick demonstration of building a machine learning model using the visual interface, showcasing the real-time model building process and the evaluation of results. The Azure Machine Learning service is presented as a full-fledged, cloud-based platform for creating, managing, and publishing ML models, with a strong emphasis on its user-friendly interface and robust capabilities.

Takeaways

  • 📚 AI is a branch of computer science that simulates human intelligence, while machine learning is a subset of AI that involves teaching software to make predictions based on data.
  • 🛠️ Azure Machine Learning is a key service for building machine learning models, which includes training, packaging, validating, and deploying models as web services.
  • 📊 The process of building a machine learning model involves training, validating, and deploying the model, followed by monitoring and retraining for improved results.
  • 💻 Azure Machine Learning provides tools such as notebooks in Python or R, a visual designer for drag-and-drop model building, and automated machine learning (AutoML) for algorithm selection.
  • 🔍 AutoML allows users to test various algorithms on their data to find the best performer, which can then be deployed as a web service.
  • 📈 Azure Machine Learning Studio is a web-based interface for managing the entire Azure Machine Learning service, including notebooks, Automated ML, and the visual designer.
  • 🧠 The Azure Machine Learning workspace is a top-level resource that brings together compute resources, permissions, experiments, pipelines, and model deployments.
  • 🔧 Users can manage compute resources and data stores such as Azure Blob Storage and Azure File Share through Azure Machine Learning.
  • 📝 The Designer feature in Azure Machine Learning allows for visual building of machine learning models with a drag-and-drop interface, without writing code.
  • 🔗 Azure Machine Learning integrates with other services for asset management, including datasets, experiments, pipelines, and endpoints for model deployment.
  • ⚙️ A compute target, such as a virtual machine, is created to run the machine learning workflow, simplifying the process of model training and evaluation.

Q & A

  • What is the main focus of the discussed episode?

    -The main focus of the discussed episode is on Azure Artificial Intelligence (AI) Services, specifically Machine Learning Studio and Machine Learning Service.

  • What is the general definition of AI?

    -AI is a branch of computer science where software is used to simulate human intelligence and capabilities.

  • How is Machine Learning related to AI?

    -Machine Learning is a subcategory of AI where software is taught to draw conclusions and make predictions based on data.

  • What is the key service in Azure for building a Machine Learning model?

    -The key service in Azure for building a Machine Learning model is called Azure Machine Learning.

  • What are the typical steps involved in building a Machine Learning model?

    -The typical steps include training the model based on data, packaging and validating the model, and if results are satisfactory, deploying the model as a web service, monitoring, and retraining the model for better results.

  • What tools does Azure Machine Learning provide to assist in the process of building Machine Learning models?

    -Azure Machine Learning provides tools like notebooks written in Python or R, a visual designer for drag-and-drop model building, and automated machine learning (AutoML) for algorithm selection and parameter tuning.

  • What is the role of a Machine Learning workspace in Azure?

    -A Machine Learning workspace in Azure is a top-level resource that ties everything together, including compute resources, permissions, runs, pipelines, experiments, history, and connections to external services.

  • How does Automated ML (AutoML) assist in the model building process?

    -AutoML assists by allowing users to apply various algorithms to their data, tweak parameters, and identify the best performing model to deploy as a web service.

  • What is the purpose of the visual designer in Azure Machine Learning?

    -The visual designer allows users to build machine learning models using a drag-and-drop interface without writing any code.

  • What is a compute target in the context of Azure Machine Learning?

    -A compute target is a virtual machine or compute resource that is used to run machine learning workflows.

  • How can users manage their machine learning models and assets in Azure?

    -Users can manage their machine learning models and assets through the Azure Machine Learning workspace, which provides features for asset management, compute resource management, and data store connections.

  • What is the difference between the old Machine Learning Studio and the one discussed in the episode?

    -The old Machine Learning Studio is no longer actively developed, whereas the one discussed is a part of the new Azure Machine Learning service with additional features and an integrated studio experience.

Outlines

00:00

📚 Introduction to Azure AI and Machine Learning

The first paragraph introduces the viewer to the topic of artificial intelligence products within Azure. It explains that AI is a branch of computer science focused on simulating human intelligence, and machine learning is a subcategory of AI that involves teaching software to make predictions based on data. The key service in Azure for machine learning is Azure Machine Learning, which assists in building models, training, packaging, validating, and deploying these models as web services. The paragraph also touches on the tools provided by Azure Machine Learning, such as notebooks in Python or R, a visual designer for a drag-and-drop experience, and the concept of automl for automated algorithm selection. Additionally, it mentions the feature of pipelines for end-to-end model building and the Azure Machine Learning Studio for a web-based interface to manage the entire service.

05:02

🚀 Azure Machine Learning Service Overview and Workflow

The second paragraph delves into the practical aspects of using Azure Machine Learning Service. It describes the process of submitting a machine learning pipeline for execution, which involves creating a new experiment and running the pipeline, with the duration depending on the complexity and chosen machine. The paragraph emphasizes the real-time visual feedback provided by the UI during model building. For data scientists, the service offers the ability to check evaluation results, logs, outputs, and visualize datasets used for training. It summarizes the Azure Machine Learning Service as an end-to-end cloud-based platform for creating, managing, and publishing machine learning models. The Machine Learning workspace is highlighted as a top-level resource that ties together all components, including compute resources, permissions, runs, pipelines, experiments, and model deployments. The paragraph also clarifies the difference between the old 'Machine Learning Studio' and the current 'Azure Machine Learning Studio', noting that the former is no longer actively developed. Key features of the service include notebooks, automated ML, visual designer for code-free pipeline building, data and compute resource management, and integration into machine learning pipelines for orchestration of tasks.

Mindmap

Definition of AI
Definition of Machine Learning
Building a Model
AI and Machine Learning Overview
Machine Learning Studio
Machine Learning Workspace
Notebooks
Automated ML (AutoML)
Visual Designer
Pipelines
Components
Training
Packaging and Validating
Deployment
Monitoring and Retraining
Process of Model Development
Asset Management
Compute Resources Management
Data Stores
Azure Machine Learning Service
Navigation to Resource Group
Launching Studio
Creating Compute Target
Building Workflow
Training and Scoring Model
Submitting Pipeline
Evaluation and Visualization
Demonstration of Azure Machine Learning Studio
Old Machine Learning Studio
New Machine Learning Service
Service Evolution and Transition
Upcoming Topics
Azure Artificial Intelligence (AI) Services
Alert

Keywords

💡Artificial Intelligence (AI)

Artificial Intelligence (AI) is a branch of computer science that aims to create systems capable of performing tasks that would typically require human intelligence. In the context of the video, AI is used to simulate human capabilities, such as learning and problem-solving, using software. It is the overarching theme that encompasses the various services and tools discussed, like Machine Learning and Azure Machine Learning.

💡Machine Learning

Machine Learning is a subcategory of AI that focuses on the development of algorithms that allow software to learn from and make predictions or decisions based on data. It is central to the video's content, as it is the process of teaching software to draw conclusions, which is facilitated by Azure Machine Learning services.

💡Azure Machine Learning

Azure Machine Learning is a cloud-based platform provided by Microsoft that assists in building, training, deploying, and managing machine learning models. It is key to the video's narrative as it offers a suite of tools and services that streamline the machine learning process, from data preparation to model deployment.

💡Model Training

Model Training refers to the process of teaching a machine learning model to make predictions or decisions based on a dataset. It is a fundamental step in the machine learning lifecycle and is highlighted in the video as a core component of what Azure Machine Learning helps automate and simplify.

💡Web Services

In the context of the video, Web Services are the applications that are deployed after a machine learning model has been trained and validated. They allow the model to be accessed and used over the internet, making the predictive capabilities of the model available to other applications or users.

💡Automated ML (AutoML)

Automated ML, or AutoML, is a feature of Azure Machine Learning that automates the process of selecting the best machine learning model by testing various algorithms and parameters. It is showcased in the video as a time-saving tool that allows data scientists to quickly identify the most effective model for their data.

💡Pipelines

Pipelines in Azure Machine Learning are end-to-end workflows that automate the machine learning process from data preparation to model deployment. They are depicted in the video as a way to streamline and manage the entire machine learning lifecycle.

💡Machine Learning Studio

Machine Learning Studio is a web-based visual interface within Azure Machine Learning that allows for the management and execution of machine learning workflows. It is used in the video to demonstrate how users can interact with Azure Machine Learning services through a user-friendly interface.

💡Compute Resources

Compute Resources in the video refer to the virtual machines and processing power provided by Azure to train, package, validate, and deploy machine learning models. They are an essential part of the infrastructure that supports the machine learning process.

💡Asset Management

Asset Management in the context of Azure Machine Learning involves organizing and managing the various components of a machine learning project, such as datasets, experiments, models, and endpoints. It is emphasized in the video as a critical feature for maintaining and tracking the progress and components of machine learning projects.

💡Linear Regression

Linear Regression is a statistical method used in machine learning to model the relationship between a dependent variable and one or more independent variables. In the video, it is used as an example of how a machine learning model can be trained to forecast prices.

Highlights

AZ-900 Episode 16 focuses on Azure Artificial Intelligence (AI) Services, specifically Machine Learning Studio & Service.

AI is defined as a branch of computer science that simulates human intelligence and capabilities with software.

Machine Learning is a subcategory of AI where software is taught to make predictions based on data.

Azure Machine Learning is a key service for building machine learning models.

The process of building a machine learning model includes training, packaging, validating, and deploying as web services.

Azure Machine Learning provides tools such as notebooks in Python or R and a visual designer for a drag-and-drop experience.

Automated Machine Learning (AutoML) allows for the testing of various algorithms on data to find the best performer.

Pipelines feature in Azure Machine Learning enables end-to-end building of machine learning models.

Azure Machine Learning Studio is a web-based visual interface for managing the entire Azure Machine Learning service.

Notebooks allow users to create scripts or try out Microsoft-provided samples for building machine learning models.

AutomatedML simplifies the process of selecting the best algorithm for a given dataset and deploying it as a web service.

The Designer in Azure Machine Learning offers a visual, drag-and-drop method for building machine learning models without coding.

Asset management in Azure Machine Learning includes datasets, experiments, pipelines, models, and endpoints.

Compute resources and data stores like Azure Blob Storage and Azure File Share are managed within Azure Machine Learning.

A machine learning model can be trained using a simple workflow with drag-and-drop data selection and cleaning steps.

Linear regression is used in the example to forecast prices, showcasing the ease of model training in Azure.

Experiments can be created and submitted to run, with the UI showing real-time model building progress.

Data scientists can evaluate results, check logs, and visualize datasets for further analysis after model training.

Azure Machine Learning Service is an end-to-end cloud-based platform for creating, managing, and publishing machine learning models.

Machine Learning Workspace is a top-level resource that ties together all components of the Azure Machine Learning service.

The old Machine Learning Studio is different and no longer actively developed; new customers are encouraged to use the new Azure Machine Learning.

Key features of Azure Machine Learning Service include notebooks, Automated ML, a visual designer, and integrated pipelines for model management.