DefinePK

Abstract

Knowledge Graph Reasoning (KGR) is a rapidly expanding field of study that aims to infer new facts from preexisting facts using the mining logic rules that underlie knowledge graphs (KGs). The use of KGs in numerous AI applications, including question-answering and recommendation systems, has been shown to greatly benefit from it. The present KGR models can be broadly classified into three categories based on the graph types: temporal models, multi-modal models, and static models. The majority of early research in this field concentrated on static KGR, while more recent studies have attempted to use temporal and multi-modal information, which is more useful and applicable to real-world scenarios. Nevertheless, there aren’t any survey articles or open-source resources that provide a thorough summary and discussion of models in this crucial area. We initially do a survey for knowledge graph reasoning tracing from static to temporal and finally to multi-modal KGs in order to close the gap. In specifics, bi-level taxonomy—that is, top-level (graph kinds) and base-level (techniques and scenarios)—is used to evaluate the models. Additionally, a summary and presentation of the datasets and performances are provided.