GCMIPA is under review in TKDE 2023.
The related datasets can be downloaded in NELL-995 FB15k-237, then, put them into "/datasets".
To reproduce the results, the hyperparameters are set in model.py
.
python model.py NELL-995 concept_organizationheadquarteredincity 10 10
RotatE HAKE PRA DeepPath MINERVA CURL GRAIL NBFNet PathCon SQUIRE
@article{DBLP:journals/tkde/ZhuoWZW24,
author = {Xingrui Zhuo and
Gongqing Wu and
Zan Zhang and
Xindong Wu},
title = {Geometric-Contextual Mutual Infomax Path Aggregation for Relation
Reasoning on Knowledge Graph},
journal = {{IEEE} Trans. Knowl. Data Eng.},
volume = {36},
number = {7},
pages = {3076--3090},
year = {2024},
url = {https://doi.org/10.1109/TKDE.2024.3360258},
doi = {10.1109/TKDE.2024.3360258},
timestamp = {Tue, 18 Jun 2024 09:25:13 +0200},
biburl = {https://dblp.org/rec/journals/tkde/ZhuoWZW24.bib},
bibsource = {dblp computer science bibliography, https://dblp.org}
}