KGRL: An OWL2 RL Reasoning System for Large Scale Knowledge Graph

• Published in: 2016 12th International Conference on Semantics, Knowledge and Grids (SKG) pp. 83 - 89
• Main Authors: Yuze Wei, Jie Luo, Huiyuan Xie
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-08-2016
• Subjects: Algorithm design and analysis; Cognition; Inference algorithms; Knowledge Graph; Ontologies; Optimization; OWL2 RL; Reasoning; Resource description framework; Sparks
• DOI: 10.1109/SKG.2016.020

Currently knowledge graph has been widely applied to various fields. Although the data scale is large, there is still a lot of useful but implicit information in it. Thus, a powerful reasoning system is required to derive these data. However, current reasoning systems cannot accomplish this task very well. On the one hand, stand-alone reasoning systems cannot meet the demand of large data. On the other hand, the reasoning ability of existing distributed reasoning systems is limited because of the lack of expressive inference rules. In this paper, we propose and implement a distributed reasoning system KGRL for knowledge graph based on OWL2 RL. It has a more powerful reasoning ability due to more expressive rules. It also supports optimization for redundant data. Besides, a rule-based algorithm is designed to find the inconsistent data. Experimental results show that KGRL can derive more implicit information efficiently compared to other reasoning systems. Moreover, KGRL is capable of eliminating redundant data, which can reduce the storage of knowledge graph by an average of 42%. Finally, KGRL also performs well for the detection of inconsistencies in knowledge graph.