Codes and datasets for "Improving Knowledge Graph Embedding Using Simple Constraints" (ACL-2018)

The repository provides java implementations and DB100K dataset used for the paper:

• Improving Knowledge Graph Embedding Using Simple Constraints. Boyang Ding, Quan Wang, Bin Wang and Li Guo. ACL 2018.

As well as the implementations for the following papers:

• Complex Embeddings for Simple Link Prediction. Théo Trouillon, Johannes Welbl, Sebastian Riedel, Éric Gaussier and Guillaume Bouchard. ICML 2016.
• Regularizing Knowledge Graph Embeddings via Equivalence and Inversion Axioms. Pasquale Minervini, Luca Costabello, Emir Muñoz, Vít Nováček and Pierre-Yves Vandenbussche. ECML 2017.

Datasets we used are in the corresponding subfolder contained in datasets/ with the following formats:

• _train.txt,_valid.txt,_test.txt; training, valid, test set with string id; format: e1\tr\te2\n
• _cons.txt; approximate entailment constraints; formant: r1,r2\tconfidence\n, where '-' denotes the inversion

python data.py data_folder

java -jar -train data_folder/train.txt -valid data_folder/valid.txt -test data_folder/test.txt

You can changes parameter when training the model

k = number of dimensions
lmbda = L2 regularization coffecient
neg = number of negative samples
mu = AER regularization coffecient

@inproceedings{boyang2018:aer,
	author = {Ding, Boyang and Wang, Quan and Wang, Bin and Guo, Li},
	booktitle = {56th Annual Meeting of the Association for Computational Linguistics},
	title = {Improving Knowledge Graph Embedding Using Simple Constraints},
	year = {2018}
}

For all remarks or questions please contact Quan Wang: wangquan (at) iie (dot) ac (dot) cn .