AI-102 Question 416
Select 4You are tasked with automating the processing of invoices for a multinational company. The invoices vary by region and format, but they all share common fields such as Invoice Number, Date, and Total Amount. You decide to use Azure Form Recognizer to create a composed document intelligence model. What steps must you take to build and deploy this model?
- A
Train multiple custom models for different invoice formats and combine them into a composed model.
- B
Use a prebuilt model provided by Azure Form Recognizer without any custom training.
- C
Manually label a training dataset for each invoice format to extract the required fields.
- D
Test the composed model using sample invoices to verify its accuracy.
- E
Deploy the composed model to an Azure resource for production use.
Show answer and explanation
Correct answers: A, C, D, E
Explanation
To create a composed document intelligence model in Azure Form Recognizer, you need to train multiple custom models tailored to different document formats or layouts. These models are then combined into a composed model to handle diverse input documents. Testing the model with sample data ensures it performs as expected, and deploying it to an Azure resource makes it accessible for production use. Prebuilt models are not suitable here because the scenario requires extracting custom fields from various formats.
- A. Correct.
Correct. A composed model allows you to combine multiple custom models trained for different document types or formats, making it suitable for handling varying invoice formats.
- B. Incorrect.
Incorrect. While prebuilt models may work for some scenarios, they are not tailored to specific document layouts, and this scenario requires extracting custom fields from various invoice formats.
- C. Correct.
Correct. Manually labeling a training dataset is necessary to train custom models for specific fields like Invoice Number, Date, and Total Amount.
- D. Correct.
Correct. Testing the composed model with sample invoices ensures its accuracy and reliability before deploying it to a production environment.
- E. Correct.
Correct. Deploying the composed model to an Azure resource, such as an endpoint, allows it to be used in production for automated invoice processing.