StarGraph: Knowledge Representation Learning based on Incomplete Two-hop Subgraph

27 May 2022  ·  Hongzhu Li, Xiangrui Gao, Linhui Feng, Yafeng Deng, Yuhui Yin ·

Conventional representation learning algorithms for knowledge graphs (KG) map each entity to a unique embedding vector, ignoring the rich information contained in the neighborhood. We propose a method named StarGraph, which gives a novel way to utilize the neighborhood information for large-scale knowledge graphs to obtain entity representations. An incomplete two-hop neighborhood subgraph for each target node is at first generated, then processed by a modified self-attention network to obtain the entity representation, which is used to replace the entity embedding in conventional methods. We achieved SOTA performance on ogbl-wikikg2 and got competitive results on fb15k-237. The experimental results proves that StarGraph is efficient in parameters, and the improvement made on ogbl-wikikg2 demonstrates its great effectiveness of representation learning on large-scale knowledge graphs. The code is now available at \url{}.

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Link Property Prediction ogbl-wikikg2 StarGraph + TripleRE Validation MRR 0.7288 ± 0.0008 # 5
Test MRR 0.7201 ± 0.0011 # 5
Number of params 86762146 # 21
Ext. data No # 1


No methods listed for this paper. Add relevant methods here