I am working on a use case where we are having 5 types of document and we need to extract the same fields from these documents.
We are planning to use AI Center and DU. We can achieve it for these 5 types of documents.
However there is a scenario where other documents type can also come but we don’t have any samples. Those are totally dynamic. Can we use communication mining to extract those same fields from those new dynamic documents?
@Ritaman_Baral Communication mining is not used to extracting data from files please look into this
you can use Document Understanding with Machine Learning
in both scenarios there is a ML only working to extract
you can feed in files…but this use case does not fit
instead based on how dynamic it is try to create a skill from already existing models and check
if that does not work then may be you can use openai connectors to send the data and ask to extract fields
cheers
We can’t use openai since this is a bank !!
Then try to train a model with as many samples as possible …then you should atleast get 90% accuracy that does the job
Cheers
Then I would create 6 document types?
5 for the ones which I have documents and the other for the dynamic ones right ?
Yes…if the new ones dall into any bucket they would be extracted and remaining would go into new bucket and you can uplod those new ones also and retrain the model if you want to increase by retrining
Cheers
This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.