search

migalkin / StarE

EMNLP 2020: Message Passing for Hyper-Relational Knowledge Graphs
MIT License
87 stars 17 forks source link

How to define main triple in JF17k #1

Closed PlusRoss closed 3 years ago

PlusRoss commented 3 years ago

Hi,

Thanks for the amazing work! May I ask how do you define the main triple and qualifiers in dataset JF17k? It seems that this dataset is unlike the other two datasets where the main triple and qualifiers are already provided as part of the data format.

migalkin commented 3 years ago

Hi, indeed, originally, JF17K resembles hypergraphs with one relation over several entities. In our work, we used JF17K processed by Rosso et al - the paper describes the methodology. The HINGE repo is available on Github

PlusRoss commented 3 years ago

Thanks for your prompt reply! However, I checked the paper and github repo but did not find how they transform the data. I may miss some information. If you know the methodology or the location of description in that paper, could you tell me it?

migalkin commented 3 years ago

Sorry for the delayed reply. I traced the citation history and believe the conversion procedure (hypergraph JF17K -> hyper-relational JF17K) stems from the work of Guan et al which is NaLP. Their github repo defines a conversion script, it might be what you are looking for.

Thanks!

PlusRoss commented 3 years ago

Thanks for your reply. It solves my problem!