Knowledge Graphs

Abstract

In this paper we provide a comprehensive introduction to knowledge graphs, which have recently garnered significant attention from both industry and academia in scenarios that require exploiting diverse, dynamic, large-scale collections of data. After some opening remarks, we motivate and contrast various graph-based data models and query languages that are used for knowledge graphs. We discuss the roles of schema, identity, and context in knowledge graphs. We explain how knowledge can be represented and extracted using a combination of deductive and inductive techniques. We summarise methods for the creation, enrichment, quality assessment, refinement, and publication of knowledge graphs. We provide an overview of prominent open knowledge graphs and enterprise knowledge graphs, their applications, and how they use the aforementioned techniques. We conclude with high-level future research directions for knowledge graphs.

KG4Vis: A Knowledge Graph-Based Approach for Visualization Recommendation

“…According to prior studies [17,43], KGE methods can mainly be categorized into 3 classes: translational models, tensor decomposition models and neural models. Translational models aim to model the relations between two entities as a translation in space.…”

“…RotatE [38] employs rotation in the complex space to represent relations between entities. Another class of KGE models aims to extract the embeddings of entities by applying tensor decomposition to model the graph structures [17], for example, RESCAL [32] and DistMult [53]. Recently, many neural models have also been proposed for KGE including SME [3] and ConvE [12].…”

“…Knowledge graph aims to build a structured form of human knowledge and has drawn great research attention from both the artificial intelligence (AI) research community and the industry [17,20,43]. However, little research has been done on building knowledge graphs for modeling the visualization design knowledge and common practices in real visualization applications.…”

Extracting and Inferring Personal Attributes from Dialogue

“…These personal attributes, represented in the form of knowledge graph triples (e.g. I, has hobby, volunteer) can represent large numbers of personal attributes in an interpretable manner, facilitating its usage by weaklycoupled downstream dialogue tasks (Li et al 2014;Qian et al 2018;Zheng et al 2020a,b;Hogan et al 2021).…”

“…Personalization across a wide range of tasks (recommending food, movies and music by multi-task dialogue agents such as Alexa, Siri and Assistant) however can require orders of magnitude more personal attribute features. This makes KG triples ideal for representing them, given the advantages of this data structure in selecting and utilizing pertinent features (Li et al 2014;Hogan et al 2021).…”