Master the workflow for creating custom document intelligence models using Azure AI. From Docker setup to training, learn the exact steps to excel in your AI-102 certification and streamline document automation.
Table of Contents
Question
Xerigon Corporation wants to use Document Intelligence to recognize and extract information from property deeds for properties that Xerigon owns and easement documents that affect those properties. The documents are not all in the same format. You need to develop a solution that must adhere to the following:
- Automate the extraction of data from the documents, reducing labor costs and minimizing errors.
- Ensure sensitive information is handled according to Xerigon’s compliance and security requirements.
- Handle large volumes of documents quickly and efficiently.
- Tailor extraction of information that is unique to Xerigon’s business.
- Create a custom table extractor by using the Sample Labeling tool.
What should you do?
Place the choices in the correct order.
Unordered Choices:
- Create a new project.
- Train the model.
- Label the data from the documents.
- Connect to the Sample Labeling tool.
- Organize the training documents into subfolders based on common format.
- Install Docker, retrieve the Sample Labeling tool container, and run the container.
Answer
Correct Order:
- Install Docker, retrieve the Sample Labeling tool container, and run the container.
- Organize the training documents into subfolders based on common format.
- Connect to the Sample Labeling tool.
- Create a new project.
- Label the data from the documents.
- Train the model.
Explanation
You would first install Docker in order to use the Sample Labeling tool. You would then use the docker pull command to retrieve the Sample Labeling tool container:
docker pull mcr.microsoft.com/azure-cognitive-services/custom-form/labeltool:latest-2.1
Once you have retrieved the container, you would run it with the following Docker command:
docker run -it -p 3000:80 mcr.microsoft.com/azure-cognitive-services/custom-form/labeltool:latest-2.1 eula=accept
You would set up the input data before connecting to the Sample Labeling tool. It would be ideal if all your training documents were in the same format. In this scenario, the documents are in different formats. When your forms are in multiple formats, it is best to organize them into subfolders based on the format. When the time comes to train the model, you can direct the API to the correct subfolder. You would enable cross-origin resource sharing (CORS) for more flexibility. CORS helps if the web application you are using requests resources from a different domain, protocol, and port.
You would then use the Sample Labeling tool to connect to a source that contains your uploaded documents and forms and the target that contains the created labels. You can connect to the Sample Labeling tool via a web browser with the following address:
http://localhost:3000
When you create the connection, you will need to specify the shared access signature (SAS) URL of your Azure Blob Storage container which would look similar to the following:
https://<storage account>.blob.core.windows.net/<container name>?<SAS value>
The following is a graphic representation of the connection to the Blob storage account.
You would then create a new project in the Sample Labeling tool which will store the configurations and settings that consist of the following:
- The Azure Blob Storage connection
- The folder path for your source forms in the blob container
- Security token
- Document Intelligence endpoint URL
- Document Intelligence key
In the Sample Labeling tool, you would identify text and tables, and then apply labels to the text. Once all the labels and tags have been applied, you can begin training your model.
When you train your model, the model will have an ID and an average accuracy.
Microsoft Azure AI Engineer Associate AI-102 certification exam practice question and answer (Q&A) dump with detail explanation and reference available free, helpful to pass the Microsoft Azure AI Engineer Associate AI-102 exam and earn Microsoft Azure AI Engineer Associate AI-102 certification.