How to preprocess the WebQuestion Dataset? #6
Description
Hello, Thank you for providing such an excellent paper.
I noticed that your preprocessing on WebQuestion datasets are described in the paper as:
We converted the SPARQL query to return a subgraph, and used the same preprocessing steps and data splits as the existing work.
When I turned to the previous work for the implement of preprocessing, I noticed that there is something different (such as they used ID representation instead of text representation). Is there any open source code for the method of preprocessing used in the paper? I want to try some changes on it. Thanks a lot for your reply!