Publication - Paper

Efficient Parallel Translating Embedding For Knowledge Graphs

ParTrans-X Architecture


    2017/01 - 2017/03


    Denghui Zhang, Manling Li, Yantao Jia, Yuanzhuo Wang, Xueqi Cheng

Sponsored by

    National Natural Science Foundation of China (NSFC) No. 61572473



    Knowledge graph embedding aims to embed entities and relations of knowledge graphs into low-dimensional vector spaces. Translating embedding methods regard relations as the translation from head entities to tail entities, which achieve the state-of-the-art results among knowledge graph embedding methods. However, a major limitation of these methods is the time consuming training process, which may take several days or even weeks for large knowledge graphs, and result in great difficulty in practical applications.


    In this paper, we propose an efficient parallel framework for translating embedding methods, called ParTrans-X, which enables the methods to be paralleled without lock by utilizing the distinguished structures of knowledge graphs.

My Work

  • Modeled the training data of knowledge graph as hypergraph.
  • Came up with the idea to prove the validity of ParTransX by exploring the law of collisions emerging.
  • Participated in experimenting ParTransX framework to demonstrate its superiority.
  • Applied ParTransX successfully in a project of Huawei Inc. (Detail information is confidential for commercial reasons).
  • Co-authored paper was accepted as a reguler paper in WI 2017.