Universal Preprocessing Operators for Embedding Knowledge Graphs with Literals

TL;DR

This paper introduces universal preprocessing operators that transform knowledge graphs with various literal types, enabling their embedding with any method, demonstrated through promising results on multiple datasets and embedding techniques.

Contribution

It proposes a versatile set of preprocessing operators for KGs with literals, compatible with any embedding method, addressing a gap in existing modality-specific approaches.

Findings

Effective transformation of KGs with literals across multiple modalities.

Compatibility of preprocessing operators with various embedding methods.

Promising results on the kgbench dataset with different embedding techniques.

Abstract

Knowledge graph embeddings are dense numerical representations of entities in a knowledge graph (KG). While the majority of approaches concentrate only on relational information, i.e., relations between entities, fewer approaches exist which also take information about literal values (e.g., textual descriptions or numerical information) into account. Those which exist are typically tailored towards a particular modality of literal and a particular embedding method. In this paper, we propose a set of universal preprocessing operators which can be used to transform KGs with literals for numerical, temporal, textual, and image information, so that the transformed KGs can be embedded with any method. The results on the kgbench dataset with three different embedding methods show promising results.