Creating an Azure AI resource is an essential first step to designing and implementing a Microsoft Azure AI solution. This process essentially sets the stage for all your subsequent AI work within the Azure ecosystem. In this post, we’ll go through the steps required to setup an AI resource in Azure, helping you prepare for the AI-102 exam.
An AI resource in Azure could be one of several different items: A cognitive service, a machine learning workspace, a bot service, or a combination of those.
Creating a Cognitive Services Resource
Microsoft Azure’s Cognitive Services are AI services and cognitive APIs that help you create applications that are more intelligent, engaging, and discoverable. To create a cognitive service resource, follow these steps:
- Sign into the Azure portal.
- Choose Create a resource > AI + Machine Learning > Cognitive Services.
-
On the Create pane, you must fill in the required fields. These include:
- Subscription: Choose one of your available Azure subscriptions.
- Resource Group: The resource group in which to place the new components.
- Region: The geographical area in which your resource resides.
- Name: A unique name for your Cognitive Service.
- Pricing tier: Choose the pricing tier that best fits your needs.
- Review the notice about charges, then select Review + create.
- After Azure validates your settings, select Create to create the Cognitive Services resource.
Creating a Machine Learning Resource
Azure Machine Learning is a versatile cloud-based, scalable service that comes with built-in Azure resource management. To create an Azure Machine Learning resource:
- Sign into the Azure portal.
- Click on “Create a resource”, then select “Machine Learning”.
-
You’ll need to provide the following details:
- Workspace Name: A unique name to identify your workspace.
- Subscription: Select the Azure subscription that you want to associate with this workspace.
- Resource Group: Choose an existing resource group or create a new one.
- Location: Select the Azure location that you want to use.
- Storage Account: Define the Azure Storage resource that the workspace will use.
- Key Vault: Select the Key Vault resource that will store secrets used by the workspace resources.
- Application Insights: Define the Application Insights resource that will collect telemetry from the workspace.
- Container Registry: If applicable, select a container registry for image deployment.
- Once all these details are provided, click the “Review+Create” button.
- After validation, click “Create” to create the resource.
Creating a Bot Services Resource
Azure Bot Service is Microsoft’s artificial intelligence (AI) chatbot offered as a service on the Azure cloud service marketplace. To create a Bot Service resource:
- Sign into the Azure portal.
- Click on “Create a resource” then choose “AI + Machine Learning” and select “Web App Bot.”
- Now, you need to provide details like the Bot name, subscription, resource group, location, pricing tier, the bot template to use, LUIS app location, and App service plan/Location.
- After you’ve filled in these details, hit “Create”. After a few moments, your bot service resource will be ready.
Once you’ve created these AI resources, you can manage them in Azure. Understanding how these resources fit together to create your AI solution is crucial for the AI-102 exam. By mastering the creation and management of Azure AI resources, you’re setting a strong foundation for both your exam preparation and your future work in Azure AI.
Practice Test
True or False: Azure Machine Learning is a cloud-based service offered by Microsoft.
- True
- False
Answer: True
Explanation: Azure Machine Learning is a cloud-based service that is designed to aid developers and data scientists with machine learning workflows.
Which of the following services are part of Azure AI?
- a) Azure Bot Service
- b) Azure Machine Learning Service
- c) Azure Cognitive Services
- d) All of the above
Answer: d) All of the above
Explanation: Azure AI encompasses several services including Azure Bot Service, Azure Machine Learning Service, and Azure Cognitive Services.
Azure AI resources can be created through which of the following methods?
- a) Azure portal
- b) Azure CLI
- c) Azure PowerShell
- d) All of the above
Answer: d) All of the above
Explanation: Azure AI resources can be created through the Azure portal, Azure CLI, or Azure PowerShell, providing flexibility depending on user familiarity and preference.
True or False: Azure Cognitive Search can use AI capabilities to extract meaningful information and build a search index.
- True
- False
Answer: True
Explanation: Azure Cognitive Search is an AI-based cloud search service for mobile and web app development that can extract and index meaningful information.
In order to create an AI resource, do you need to have an active Azure subscription?
- Yes
- No
Answer: Yes
Explanation: To create an AI resource in Microsoft Azure, an active Azure subscription is required.
Which of the following are required to deploy Azure Machine Learning models?
- a) Workspace
- b) Compute resources
- c) Datastore
- d) All of the above
Answer: d) All of the above
Explanation: All these things are needed to deploy Azure Machine Learning models – Workspace to hold all the resources, compute resources to process data, and datastore to store datasets.
True or False: In terms of Azure Cognitive Services, all services are generally available and there are no preview services.
- True
- False
Answer: False
Explanation: In Azure Cognitive Services, there are many services that are generally available, and also some are still in the preview state.
The Azure Machine Learning designer does NOT support which of the following tasks?
- a) Data transformation
- b) Model training
- c) Monitoring the weather
- d) Model evaluation
Answer: c) Monitoring the weather
Explanation: The Azure Machine Learning designer is mainly used for building machine learning pipelines involving tasks like data transformation, model training, and model evaluation. It does not have a feature for weather monitoring.
Which of the following is NOT a benefit of Azure AI?
- a) Scalability
- b) Flexibility
- c) Data Security
- d) Free cost for all services
Answer: d) Free cost for all services
Explanation: While Azure AI does offer a lot of benefits like scalability, flexibility, and data security, it is not free of cost for all services.
True or False: It is compulsory to have programming skills for using Azure Machine Learning Studio.
- True
- False
Answer: False
Explanation: Azure Machine Learning Studio has a drag-and-drop interface that allows users with no programming skills to create and deploy models.
Interview Questions
What is the first step in creating an Azure AI resource?
The first step in creating an Azure AI resource is to sign into the Azure portal.
How do you create a new resource in Azure?
You create a new resource in Azure by clicking on the “Create a resource” button in the Azure portal, search for the desired AI service, and then follow the prompts to configure and deploy the resource.
Which Azure AI service would you use for text Analytics?
For text analytics, you would use the Azure Text Analytics API, which is part of Azure’s Cognitive Services offering.
What is the Azure Machine Learning studio used for?
The Azure Machine Learning studio is a visual interface used for building, training, and deploying machine learning models.
Where can you find the keys and endpoint for an AI resource once it is deployed on Azure?
The keys and endpoint for an AI resource once it is deployed on Azure can be found by navigating to the resource and selecting the “Keys and Endpoint” option from the left-hand menu.
Does Azure AI offer real-time speech recognition services?
Yes, Azure AI offers real-time speech recognition through the Speech service which is a part of Azure Cognitive Services.
What kind of services does Azure Cognitive Services provide?
Azure Cognitive Services provide pre-built AI services including vision, speech, language, decision, and web search services that developers can use in their applications.
How do you secure Azure AI resources?
Azure AI resources can be secured by keeping all keys and endpoints private, implementing Azure RBAC (Role-Based Access Control), setting up network security controls such as Azure Firewall or Virtual Network Service Endpoint, and by monitoring and auditing activity with Azure Monitor and Azure Activity Log.
What is the Azure Bot Service and why is it useful in the AI context?
Azure Bot Service is an Artificial Intelligence (AI) service on Microsoft Azure used to develop intelligent, enterprise-grade bots. It allows developers to build, connect, test, and deploy bots that interact naturally with users, using a range of services like text/SMS, Skype, Teams, and others.
What is Azure Form Recognizer and how is it used?
Azure Form Recognizer is an AI service that uses machine learning to identify and extract key-value pairs and table data from form documents. It is used to automate and streamline the data entry process in applications.
What types of virtual machines can be provisioned in Azure Machine Learning?
In Azure Machine Learning, one can provision CPU and GPU enabled virtual machines.
Which Azure AI service can help with personalized recommendations?
The Azure Personalizer service can help with personalized recommendations, it uses reinforcement learning in a real-time environment to provide personalized user experiences.
Is it possible to use Azure Logic Apps with Azure AI?
Yes, Azure Logic Apps can be used with Azure AI for business process automation and work-flow design via various connectors which include cognitive services as well.
What is the purpose of Azure Custom Vision and where can it be applied?
Azure Custom Vision is a service under Azure Cognitive Services which enables developers to build and refine custom image classification models. It can be applied in numerous scenarios such as identifying company logos, detecting product defects, classifying images based on specific requirements, etc.
How can Azure Databricks be integrated with Azure Machine Learning?
Azure Databricks can be integrated with Azure Machine Learning by creating an Azure ML workspace, configuring an Azure Databricks workspace, and then connecting the two. This allows for tracking of runs, deployment of models, and maintaining a consistent data science workspace.
extutorials.com