I think the idea of randomly converting any paragraph of text into different entities in a document is really genius, but I have encountered at least the following two problems in practice:
- If I want to convert its entity type., it will disappear in this document.
- I can’t convert or delete entities in documents directly.
I understand that the reason why documents can be so clever to convert their entity types is that we should not have too much mental burden when recording at the beginning, but we can convert casually in the later period, but because of his two problems, the burden is still a bit heavy when recording, so we need to think carefully about which stage it belongs to, and we can’t go back on our word.