The experimental results are different between the original paper and GitHub

AnselCmy / MetaR

Source code for the EMNLP 2019 paper "Meta Relational Learning for Few-Shot Link Prediction in Knowledge Graphs"

Apache License 2.0

111 stars 19 forks source link

The experimental results are different between the original paper and GitHub #3

Closed YanpingWu12138 closed 2 years ago

YanpingWu12138 commented 3 years ago

Hello, AnselCmy

Can you help me understand why the experimental results in your GitHub are higher than those in your original paper? Did you make improvements after you published the paper or there is a new paper here?

AnselCmy commented 2 years ago

Thanks! We finetuned the hyper-parameters after this paper was accepted. There is no significant improvement in the method, and it's not a new paper here.