Document Understanding process inquiries

[michelle.soto (MS)]

Hello,

I am fairly new to document understanding capability and need to clarify a few inquiries.

1. We do not need Ai Center to apply DU in out process but we do need a DU api key for the Intelligent Keyword Classifier trainer and Form Extractor, is this API key available in community edition?
2. Is Intelligent Keyword Classifier trainer the only trainer available without AI Center?
3. Are Form Extractor and Regex the only extractors available without AI Center?
4. My understanding is that we do not have to train the model each time we run it, if this is true, how do we apply the previously trained model to our current process?

Thank you.

[sonaliaggarwal47 (Sonali Aggarwal)]

Hi @michelle.soto

Yes, DU api key is avaiable for community users however its ai units may differe from enterprise version.

Also, Please see if below info helps,

AI Center - AI Units

[Nikhil_Zambare1 (Nikhil Zambare)]

Dear @michelle.soto,

1. We do not need Ai Center to apply DU in out process but we do need a DU api key for the Intelligent Keyword Classifier trainer and Form Extractor, is this API key available in community edition?
   Answer: Yes, UiPath provide modern DU which does not required AI center.
   But to train classiffier or trainer you need to have DU API key.

API Key:

Modern DU:

Question: Is Intelligent Keyword Classifier trainer the only trainer available without AI Center?
Answer: No, Keywordbased trainer is also there.

Question: Are Form Extractor and Regex the only extractors available without AI Center?
Answer: Yes

Question: My understanding is that we do not have to train the model each time we run it, if this is true, how do we apply the previously trained model to our current process?
Answer: This is based on your usecase. Check the model score, if its average then you have to train the model. Re-training always does not gurantee that the model will perfrom well. It might decrese the accuracy also.
It good choice to train model with lager amount of dataset which coveres every aspect of docuemnt fields.

To auto train your ML model you can make use of below:

1. Auto export dataset from document manager:
   

   

2. Auto run your train pipeline. Note that the pipeline should run after export done. Schdule it properly.

3. Retraining utilize more AI units.

Mark as solution if it solves your queries

Br,
NikZ

[michelle.soto (MS)]

Thank you, it was very helpful