Abstract

Discovering precise and specific rules from knowledge graphs is regarded asan essential challenge, which can improve the performances of many downstreamtasks and even provide new ways to approach some Natural Language Processingresearch topics. In this paper, we provide a fundamental theory for knowledgegraph reasoning based on ending anchored rules. Our theory provides precisereasons answering why or why not a triple is correct. Then, we implement ourtheory by what we called the EARDict model. Results show that the EARDict modelachieves new state-of-the-art performances on benchmark knowledge graphcompletion tasks, including a Hits@10 score of 80.38 percent on WN18RR.

 

Quick Read (beta)