Using the AI Builder Feedback Loop Documents to Tag and Train Model Using Power Apps

Like (1) Share

Report

Posted on 15 Jul 2024 12:49:56 by kkeyanb

Hi,

We have built a power automate solution that uses custom Document Extraction (Invoice) Model to extract invoice data and use the feedback loop to save Invoices that are different or not fully recognisable by the custom model we have trained.Now this new invoice document needs to be tagged and added to the model and the model needs to be trained and published.

Can we do this using a Power APP where the user browses through the documents in the feedback Loop,tags it if required and then submits the document to the model for training.This process can be done manually but then that requires manual intervention and without automation of this process the feedback Loop mechanism is not of great use as without user's intervention the feedback loop documents dont get added as training data for the Model.

Any one who has used the feedback loop to train the model or what do you do with the feedback loop documents.

Regards

Karthik

All responses (1)

Answers (0)

Suggested answer

Tjan 1,087 Super User 2025 Season 1 on 26 Jul 2024 at 07:44:23

Like (0)

Report

Using the AI Builder Feedback Loop Documents to Tag and Train Model Using Power Apps

Hi, I have experience with the feedback loop and what it basically does is add specific documents to an existing collection as training document in your AI Builder model. You will afterwards find this document when editing your AI model under the documents that need to be tagged. You can hereby choose to 1. use them as additional training data (you need to tag them afterwards) or 2. not use them and delete them.

No PowerApps can be used for this since the whole process of retraining with the feedback loop document can be done within AI Builder.