Process Steps for Form Processing Model Training

Correct Answers: A, C and E

AI Builder provides several prebuilt Machine Learning cognitive models that users can use in their Power Platform products.

AI Builder model list includes models like prediction, object detection, form processor, text classification, category classification, entity extraction, etc.

If your business needs do not match the out-of-box models, you can train your models based on your data.

AI Builder Form processing model is a computer vision model that extracts text from the image.

Businesses use this model to automate the data extraction from the standard documents, like invoices or tax documents.

When you select to create a Form Processing model option from the AI Builder screen and provide the model's name, a new screen will open.

On this screen, you will follow the steps for preparing a model for training.

The three process steps are: Choose information to extract - you need to create tags for fields you want to extract from the documents.

Add collections of the documents - you need to create the collections and upload similar documents (with the same layout) to collections.

You can make up to 100 collections.

After this, you click the Analyze button and wait until the AI builder examines the documents and detects the fields and tables you identified in the previous step.

Tag documents - this is a process of teaching the model how to extract the fields and tables.

You need to tag all uploaded documents.

After you tag the documents, you can review your model's summary and push the button Train on the next step.

Option B is incorrect because Form processing preparation creates document collections by the similarity in the documents' layouts but not documents groups by the document types.

Option D is incorrect because the Select domain is an operation step for Object detection model training but not for Form processing.

For more information about AI Builder Form processing model training, please visit the below URLs:

The Form processing AI model is used to automatically extract data from structured documents like invoices, receipts, and business cards. To train this model, you need to follow certain steps, including:

1. Tag documents: The first step is to tag the documents that you want the AI model to extract data from. Tagging refers to the process of labeling each document with metadata that describes its type, purpose, and other relevant information. This helps the AI model learn which types of documents it should be looking for and what data it should be extracting.

2. Group documents by their type: Once you have tagged the documents, the next step is to group them by their type. This involves creating different collections of documents based on their format, such as invoices, receipts, and business cards. Grouping the documents by type helps the AI model learn the unique characteristics of each type of document and how to extract data from them.

3. Choose information to extract: After grouping the documents by type, you need to select the information that you want the AI model to extract from each document. This includes identifying the specific fields that the model should be looking for, such as the date, total amount, and vendor name on an invoice. You also need to determine how the model should extract this information, such as using optical character recognition (OCR) or natural language processing (NLP).

4. Select a domain: The next step is to select a domain for the AI model. A domain refers to the specific type of document or industry that the model will be trained to recognize and extract data from. For example, you might select the finance domain if you are training the model to extract data from financial documents like invoices and receipts.

5. Create document collections: Finally, you need to create collections of documents that are representative of the types of documents the AI model will encounter in the real world. This involves gathering a large number of sample documents that the model can use to learn how to recognize and extract data from different types of documents.

In summary, the three process steps for Form processing model training are tagging documents, grouping documents by their type, and choosing information to extract. These steps help the AI model learn how to recognize and extract data from structured documents like invoices, receipts, and business cards.