Differentiable learning of rules with constants in knowledge graph

Zezhong Xu, Peng Ye, Juan Li, Huajun Chen, Wen Zhang

Abstract: Highlights•Draw attention to expanding the diversity of target rules for rule mining and emphasize the importance of constants to rules.•Propose an end-to-end differentiable rule mining method that can mine rules in an efficient way.•Experimentally demonstrate the importance of rule with constants.•Design an algorithm that outputs high-quality symbolic rules with constants.