The present validation station activity

Help Something Else
1

Hello guys

the PRESENT VALIDATION STATION’s role is a little confusing!
Is it just for showing what value is extracted?
or an important phase in the classification ( It means the bot do the classification then when human do the validation he learns from that )?

thnx in advance

2

It allows an agent to validate and correct the extractions

Lets assume OCR was recognizing not fully correctly a field. Within the validation station it can be corrected.

From official docu:

The Validation Station is a user interface that can be used to validate and correct data extracted from documents using the Document Understanding framework.

3

Thank u

So it is possible to elliminate the present validation station ?
Because I have a large quantity of files that want to process with Document Understanding But because of the presence of the Validation Station, it seems an impossible job

4

ommiting the present validation station activity cuts off the posibility to manually review and correcting extraction results

from technically implementation viewpoint yes
from process / technological viewpoint - depends on how you rate the expected accurancy and the risk of failing recognitions

5

Yeah Exactly
I am trying to do the classification using Keyword-based classifier but it is failing every time it does the classification which makes the elimination of the validation part not possible :sob: :sob:
Is there a solution that makes me have a good classification score without using the validation station?

6

we implemented some flows incorporating

  • look up to reference lists for value corrections after OCRing and extraction
  • checking the different accurancy percentages and fork depending on it:
    • continue: without validation station (Background processing / Dunkelverarbeitung)
    • present it to agent within validation station

But before stepping too much into the technical details lets loop @loginerror in. Maybe the DU product team has this feature already on the roadmap and will share some previews with us

1 Like
7

thank you a lot

8

This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.

10

I asked the Document Understanding gurus and this is the reply :slight_smile:

The classification score is determined based on the classification activity and depends on the configuration as well as the documents provided by the user - it does not require the user to use the validation station, but would rather rely on the algorithm and the users configurations

You can find more details here:

11

thank you maciej
As you told me to eliminate the validation station I faced an issue
can you visit this link and if you have an idea don’t hesitate to contact me