SemReasoner - A High-Performance Knowledge Graph Store and Rule-Based Reasoner

Abstract

Knowledge graphs have become essential for integrating data from heterogeneous sources powering intelligent applications. Integrating data from various sources often results in incomplete knowledge that needs to be enriched based on custom inference rules. Handling a large number of facts requires a scalable storage layer that must be seamlessly integrated into the reasoning algorithms to guarantee efficient evaluation of rules and query answering over the knowledge graph. To this end, we present SemReasoner, a comprehensive, scalable, high-performance knowledge graph store and rule-based reasoner. SemReasoner includes a deductive reasoning engine and fully supports document store functionality for JSON documents. SemReasoner’s modular architecture is easy to extend and integrate into existing IT landscapes and applications. We evaluate SemReasoner against the state-of-the-art rule-based reasoning engines using test cases from OpenRuleBench. The results show that SemReasoner outperforms existing engines in most test cases.