Differentiable learning of rules with constants in knowledge graph

Knowledge reasoning helps overcome the incompleteness of knowledge graphs (KGs) and has significantly contributed to the development of large KGs. Rule mining, one of the key tasks of knowledge reasoning, studies the problem of learning interpretable inference patterns over KGs. Existing rule mining methods mainly focus on learning rules that consist of different relations and variables, restricting the form of rules to be closed path. While rules could be diverse and in order to enrich the forms of rules, we argue that constants should also be considered in the rule mining process. We propose an Elegant Differentiable rUle learning with Constant mEthod (EduCe).1 We propose a constant operator and dynamic weight mechanism, which choose the constants that should be added and decrease the number of parameters, respectively. The model could mine diverse and accurate rules in an efficient way with these modules. The experimental results on several knowledge graph completion benchmarks show that EduCe achieves state-of-the-art link prediction results among differentiable rule mining methods and successfully learns diverse and high-quality rules. & COPY; 2023 Elsevier B.V. All rights reserved.

Automated summary

Knowledge reasoning is crucial for overcoming the limitations of knowledge graphs (KGs) and has made significant contributions to the advancement of large KGs. Rule mining, an important task in knowledge reasoning, focuses on learning interpretable inference patterns from KGs. Existing methods mainly concentrate on closed path rules with various relations and variables, disregarding the inclusion of constants. In this paper, we propose EduCe, an Elegant Differentiable rUle learning with Constant mEthod, which considers constants in the rule mining process and incorporates a constant operator and dynamic weight mechanism to enhance rule diversity and accuracy. Experimental results on multiple knowledge graph completion benchmarks demonstrate that EduCe outperforms other differentiable rule mining methods in terms of link prediction and can effectively learn a wide range of high-quality rules.