I need to understand one thing.
Sometimes for a particular field/document the confidence is less than 100
However the data extracted is correct
Why this happens any explanation in layman’s terms ?
In document understanding confidence score indicates how sure the AI model is about the data it has extracted, based on patterns it has learned. However, even if the data is correct, the score might not be 100%
1 Like
When a field or document shows a confidence score less than 100% but still has the correct data, it means the system is uncertain about the result, even though it got it right.
Here’s a simple explanation:
The extraction tool (like UiPath’s Document Understanding) uses machine learning models to predict the right information. These models rely on patterns from the data they’ve been trained on. Sometimes, the tool may encounter variations, formatting changes, or new examples it hasn’t seen much of before. This can make the model unsure if the data it’s extracted is truly accurate, even though it is. It’s like a person who gives the right answer on a test but isn’t 100% confident about it.
So, low confidence doesn’t mean the data is wrong, just that the model is being cautious.
LLM helped me to write this but it’s validated by me.
Thanks,
Ashok 
1 Like
This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.