Question of results on relation category

[ninjaX2o]

In your paper, predicting the tail of entity 1-N is better than the head, but predicting one to many is often more difficult, and the same applies to N-1. So I suspect you wrote the results upside down

[reyingpai]

请问你解决了吗？你有出现其他数据集的实验结果和论文相差很大么

[reyingpai]