Context Graph

TL;DR

This paper introduces Context Graphs, an extension of traditional Knowledge Graphs that incorporates contextual information like time, location, and provenance, enhancing reasoning and knowledge representation capabilities.

Contribution

The paper proposes the concept of Context Graphs and a reasoning paradigm CGR$^3$ that leverages large language models for improved knowledge graph completion and question answering.

Findings

CGR$^3$ significantly improves KG completion accuracy.

CGR$^3$ enhances KG question answering performance.

Context Graphs provide richer, more nuanced knowledge representations.

Abstract

Knowledge Graphs (KGs) are foundational structures in many AI applications, representing entities and their interrelations through triples. However, triple-based KGs lack the contextual information of relational knowledge, like temporal dynamics and provenance details, which are crucial for comprehensive knowledge representation and effective reasoning. Instead, \textbf{Context Graphs} (CGs) expand upon the conventional structure by incorporating additional information such as time validity, geographic location, and source provenance. This integration provides a more nuanced and accurate understanding of knowledge, enabling KGs to offer richer insights and support more sophisticated reasoning processes. In this work, we first discuss the inherent limitations of triple-based KGs and introduce the concept of CGs, highlighting their advantages in knowledge representation and reasoning. We then present a context graph reasoning \textbf{CGR$^3$} paradigm that leverages large language models (LLMs) to retrieve candidate entities and related contexts, rank them based on the retrieved information, and reason whether sufficient information has been obtained to answer a query. Our experimental results demonstrate that CGR$^3$ significantly improves performance on KG completion (KGC) and KG question answering (KGQA) tasks, validating the effectiveness of incorporating contextual information on KG representation and reasoning.