SEEK Framework for Knowledge Graph Embeddding

Source code for the ACL 2020 paper "SEEK: Segmented Embedding of Knowledge Graphs".

Training

make && ./main -dataset DB100K -num_thread 24 -model_path seek.model

Link Prediction Task

./main -dataset DB100K -num_thread 24 -model_path seek.model -prediction 1

Triple Classification Task

./main -dataset DB100K -num_thread 24 -model_path seek.model -classification 1

Command Line Option

Option	Description
-dataset	Dataset
-num_thread	Number of threads
-embed_dim	Dimension of embeddings
-num_seg	Number of segments
-neg_sample	Negatives samples
-num_epoch	Epochs for training
-model_path	Model path
-lambda	L2 weight regularization penalty
-lr	Init learning rate

Citation

Please cite the following paper if you use this code in your work.

@inproceedings{xu-etal-2020-seek,
    title = "{SEEK}: Segmented Embedding of Knowledge Graphs",
    author = "Xu, Wentao and Zheng, Shun and He, Liang and Shao, Bin and Yin, Jian and Liu, Tie-Yan",
    booktitle = "Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics",
    month = jul,
    year = "2020",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://www.aclweb.org/anthology/2020.acl-main.358",
    pages = "3888--3897",
}