The dataset is uploaded in the zenedo: https://zenodo.org/records/13348627
This code base extracts subgraphs from an endpoint, form sparqls, extract answers. The above info would be sent to LLM for question generation.
- Python Version:3.8 and above
- pip install -r requirement.tex
-
go to folder 'extract'
-
run endExtract.py and statistics.py
-
they would extract subgraphs based on the network motifs and output json files with related info
-
go to folder generation
-
rub mistral_baseline.py
-
generate question from LLM
@incollection{yan2024bridging,
title={Bridging the Gap: Generating a Comprehensive Biomedical Knowledge Graph Question Answering Dataset},
author={Yan, Xi and Westphal, Patrick and Seliger, Jan and Usbeck, Ricardo},
booktitle={ECAI 2024},
pages={1198--1205},
year={2024},
publisher={IOS Press}
}
```****