Overview on different components of UiPath Document Understanding Framework.

UiPath Document Understanding pairs RPA and AI to automatically process your documents. It enables intelligent document processing within automation workflows, thus, allowing automation of complex and cognitive processes which are usually highly manual. Different components of Document Understanding framework is shown below.

1.LoadTaxonomy

It models the type of documents that the project will work with.

2. Digitize

If the pdf is a digital native document(whose text are directly copyable), we can skip this step. However, if the doc is with scanned images, or raster PDFs, we need OCR (Optical Character Recognition) technology to transform the image to text.

3. Classify

If we are dealing with more than one document sample( eg: need to process an electricity bill format as well as a medical chart),the robot will need to identify which document category the current case falls into, and process it accordingly. This can be done either in a simple, keyword-based manner, or, if no such rule can be defined, with the help of UiPath’s AI Center.

4. Extract

This stage identifies specific data (e.g. billed value, a name, an address etc) out of an entire document.

5. Export

To export the extracted data as per the business needs.

Note: as described in the image, both #3 and #4 steps has to go through #Train(To train the robot about each fields) and #Validate(validating the bot's output and make it more efficient as much as possible) states.