Relation Adversarial Network for Low Resource Knowledge Graph Completion

被引:44
作者
Zhang, Ningyu [1 ,2 ]
Deng, Shumin [2 ,3 ]
Sun, Zhanlin [4 ]
Chen, Jiaoyan [5 ]
Zhang, Wei [1 ,2 ]
Chen, Huajun [2 ,3 ]
机构
[1] Alibaba Grp, Hangzhou, Peoples R China
[2] AZFT Joint Lab Knowledge Engine, Hangzhou, Peoples R China
[3] Zhejiang Univ, Hangzhou, Peoples R China
[4] Carnegie Mellon Univ, Sch Comp Sci, Pittsburgh, PA 15213 USA
[5] Univ Oxford, Dept Comp Sci, Oxford, England
来源
WEB CONFERENCE 2020: PROCEEDINGS OF THE WORLD WIDE WEB CONFERENCE (WWW 2020) | 2020年
关键词
Knowledge Graphs; Low Resource Knowledge Graph Completion; Adversarial Transfer Learning; Link Prediction; Relation Extraction;
D O I
10.1145/3366423.3380089
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Knowledge Graph Completion (KGC) has been proposed to improve Knowledge Graphs by filling in missing connections via link prediction or relation extraction. One of the main difficulties for KGC is a low resource problem. Previous approaches assume sufficient training triples to learn versatile vectors for entities and relations, or a satisfactory number of labeled sentences to train a competent relation extraction model. However, low resource relations are very common in KGs, and those newly added relations often do not have many known samples for training. In this work, we aim at predicting new facts under a challenging setting where only limited training instances are available. We propose a general framework called Weighted Relation Adversarial Network, which utilizes an adversarial procedure to help adapt knowledge/features learned from high resource relations to different but related low resource relations. Specifically, the framework takes advantage of a relation discriminator to distinguish between samples from different relations, and help learn relation-invariant features more transferable from source relations to target relations. Experimental results show that the proposed approach outperforms previous methods regarding low resource settings for both link prediction and relation extraction.
引用
收藏
页码:1 / 12
页数:12
相关论文
共 60 条
  • [1] Bollacker K, 2008, P 2008 ACM SIGMOD IN, P1247, DOI DOI 10.1145/1376616.1376746
  • [2] Bordes A., 2013, ADV NEURAL INFORM PR, V26, P2787, DOI DOI 10.5555/2999792.2999923
  • [3] Cai Liwei, 2017, ARXIV PREPRINT ARXIV
  • [4] Cao Zhangjie, 2017, ARXIV 170707901
  • [5] Chen M, 2019, ARXIV190901515, P4216, DOI 10.18653/v1/D19-1431
  • [6] Re-weighted Adversarial Adaptation Network for Unsupervised Domain Adaptation
    Chen, Qingchao
    Liu, Yang
    Wang, Zhaowen
    Wassell, Ian
    Chetty, Kevin
    [J]. 2018 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2018, : 7976 - 7985
  • [7] Dettmers T, 2018, AAAI CONF ARTIF INTE, P1811
  • [8] Devlin J, 2019, 2019 CONFERENCE OF THE NORTH AMERICAN CHAPTER OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS: HUMAN LANGUAGE TECHNOLOGIES (NAACL HLT 2019), VOL. 1, P4171
  • [9] Fu Lisheng, 2017, P 8 INT JOINT C NAT, V2, P425
  • [10] Ganin Y, 2016, J MACH LEARN RES, V17