Some confusing points
xiaomingaaa opened this issue · 2 comments
Very nice work!
I'm confused as to why the performance (i.e., Hits@10) on NELL-995 reported in your article is so different from the original GraIL article. This seems to be the case with other baselines as well. Looking forward to your feedback, thank you!
Hi, thanks for your question.
As we introduce in Sec 4.1.2, we follow [53] to evaluate the ranking of target answer entity over all the negative entities, rather than 50 randomly sampled negative ones adopted in [36].
[53] Y. Zhang and Q. Yao. 2022. Knowledge Graph Reasoning with Relational Directed Graph. In TheWebConf.
[36] K. K Teru, E. Denis, and W. L Hamilton. 2020. Inductive Relation Prediction by Subgraph Reasoning. Technical Report. arXiv:1911.06962.
Thank you for your clarity, I get it!