Comment by ezhil

The extraction happens based on a schema of nodes and edges. So, let's say you have 2 docs that have related data on a company, then they both will be connected. We use Entity resolution to combine them.