Skip to content

README.md

Latest commit

 

History

History
44 lines (31 loc) · 1.69 KB

README.md

File metadata and controls

44 lines (31 loc) · 1.69 KB

Triple Feature Propagation (TFP)

This is the code of Triple Feature Propagation (TFP) introduced in our paper: "Rethinking Smoothness for Fast and Adaptable Entity Alignment Decoding", accepted by NAACL 2025 as Findings. It is the formal version of the paper "Gradient Flow of Energy: A General and Efficient Approach for Entity Alignment Decoding".arxiv

Datasets

The dataset and the embedding we processed can be downloaded at GoogleDrive. You can also use the same datasets in DATTI.

  • ent_ids_1: ids for entities in source KG;
  • ent_ids_2: ids for entities in target KG;
  • rel_ids_1: ids for relations in source KG;
  • rel_ids_2: ids for relations in target KG;
  • sup_ent_ids: training entity pairs;
  • ref_ent_ids: testing entity pairs;
  • triples_1: relation triples encoded by ids in source KG;
  • triples_2: relation triples encoded by ids in target KG;

Just run main.py

Environment

  • Python == 3.7.0
  • tensorflow == 2.6.0
  • Numpy
  • tqdm

Acknoledgement

We appreciate DATTI for their open-source contributions.

Cite

Please consider citing this paper if you use the code or data from our work. Thanks a lot ~

@inproceedings{wang2025rethinking,
  title={Rethinking Smoothness for Fast and Adaptable Entity Alignment Decoding},
  author={Wang, Yuanyi and Li, Han and Sun, Haifeng and Zhang, Lei and He, Bo and Tang, Wei and Yan, Tianhao and Qi, Qi and Wang, Jingyu},
  booktitle={Findings of the Association for Computational Linguistics: NAACL 2025},
  pages={4521--4535},
  year={2025}
}