Why should the RAPTOR function be implemented as a knowledge base as a whole?

[chg387387]

If my knowledge base contains a number of papers with different methods and conclusions, if they are mixed together, the method of paper A will be applied to the experiment of paper B, and the conclusion of paper C will be obtained, which will cause confusion in the entire knowledge system. If I change to the original method and run RAPTOR on each document separately, it will be much better.

[yingfeng]

Doc level RAPTOR is going to be recovered soon.