AI center and training robot

[Suldarie]

Hi all,

I have a process that I would like to describe very briefly. It takes invoices from a sharepoint, extracts the data (digitize document, classify document scope, data extraction scope) and waits for human validation in action center (create document validation action, wait for document validation action and resume).

I would like the corrections made by the customers in the action center to “feed” the robot in order to improve its extraction skills.

For that we have a project created in the AI center. But the robot is not really using it for now. The only thing that I have is a robot training workflow, made with a train extractor scope. But that’s it (and not sure it is connected to this AI center project).

I have to admit that I took over this project so I am not completely familiar with it.

I am a bit confused between document understanding and AI Center.

I would like to know how to train my robot with data validated from the action center.

I also have trained the robot in the AI center with around 400 invoices, but, same here, I don’t know how to implement this training inside my process.

I have tried to use the ML skill activity in my workflow but not sure how to implement that.

As you can see I am confused on how to proceed and how things work.

If you have suggestions or any tutorial that could be straightforward, that would be awesome.

[sharon.palawandram]

Hi Suldarie,

First off, you’re on the right track in designing your automation. For an invoice processing model I’m guessing you’re using a machine learning extractor in data extraction scope. Therefore you can retrain the model with human validated data through “machine learning extractor trainer”. A machine learning extractor trainer enables the collection of data that has been processed through Validation Station so that it can be imported into Data Manager. This activity can be used only within the Train Extractors Scope activity.

This methodology ensures that corrections made by the customer in action center is fed into your machine learning model which over times improves extraction capabilities.

here’s a guide on how to use machine learning trainer activity

https://docs.uipath.com/activities/docs/machine-learning-extractor-trainer

next, to distinguish between document understanding & AI center, Document Understanding is the framework that enables you to process files, digitize them and extract as you need. you can find more information on what is document understanding over here.

https://docs.uipath.com/document-understanding/docs/introduction

AI center is a service that allows you to deploy, manage, and continuously improve Machine Learning models and consume them within RPA workflows in Studio. It’s a pretty impressive service that lets you to create, train, modify & deploy out of the box & custom build machine learning models.

More information on AI center:

https://docs.uipath.com/ai-fabric/v0/docs/about-ai-center

Lastly, you can learn and train on how to build machine learning models in UiPath academy. They have very good courses that teach you to create and deploy document understanding models.

You can also find youtube videos on how to train a machine learning model

Training UiPath Document Understanding ML Models - Data Manager - Part 2 | RPA

Overall, you’re on the right track already and hope these resources help you to understand better, good luck!

[Suldarie]

Thank you so much for taking the time to reply me.

I didn’t get the chance to check this out, but I will do it for sure

Thanks again.

[sharon.palawandram]

you’re welcome!