TY - CHAP

T1 - Neuro-Symbolic Relation Extraction

AU - Yan, Xi

AU - Usmanova, Aida

AU - Möller, Cedric

AU - Westphal, Patrick

AU - Usbeck, Ricardo

PY - 2025/3/17

Y1 - 2025/3/17

N2 - Neuro-symbolic relation extraction lies at the intersection of neural networks and symbolic reasoning, presenting promising opportunities to enhance the capabilities of natural language processing (NLP) systems. Despite its potential, a comprehensive review of how these systems are developed and applied to the task of relation extraction has been lacking. This chapter addresses this gap by offering an in-depth overview of the current landscape in neuro-symbolic relation extraction, focusing on key methodologies and the datasets utilized in this field. We systematically categorize existing approaches, emphasizing how they integrate neural and symbolic components to tackle various challenges and the types of information they incorporate. Additionally, we review the datasets used to evaluate neuro-symbolic relation extraction systems, detailing their statistics, creation processes, and underlying domains. Furthermore, we discuss future research directions and challenges, such as the analysis of symbolic information and the integration of datasets with existing knowledge graphs. By synthesizing these findings, this chapter aims to provide researchers and practitioners with a clear understanding of the state of neuro-symbolic relation extraction and to inspire further innovations in this rapidly evolving field.

AB - Neuro-symbolic relation extraction lies at the intersection of neural networks and symbolic reasoning, presenting promising opportunities to enhance the capabilities of natural language processing (NLP) systems. Despite its potential, a comprehensive review of how these systems are developed and applied to the task of relation extraction has been lacking. This chapter addresses this gap by offering an in-depth overview of the current landscape in neuro-symbolic relation extraction, focusing on key methodologies and the datasets utilized in this field. We systematically categorize existing approaches, emphasizing how they integrate neural and symbolic components to tackle various challenges and the types of information they incorporate. Additionally, we review the datasets used to evaluate neuro-symbolic relation extraction systems, detailing their statistics, creation processes, and underlying domains. Furthermore, we discuss future research directions and challenges, such as the analysis of symbolic information and the integration of datasets with existing knowledge graphs. By synthesizing these findings, this chapter aims to provide researchers and practitioners with a clear understanding of the state of neuro-symbolic relation extraction and to inspire further innovations in this rapidly evolving field.

KW - Informatics

KW - Business informatics

U2 - 10.3233/faia250222

DO - 10.3233/faia250222

M3 - Chapter

SN - 978-1-64368-578-6

T3 - Frontiers in Artificial Intelligence and Applications

SP - 550

EP - 576

BT - Handbook on Neurosymbolic AI and Knowledge Graphs

A2 - Hitzler, Pascal

A2 - Dalal, Abhilekha

A2 - Mahdavinejad, Mohammad Saeid

A2 - Norouzi, Sanaz Saki

PB - IOS Press BV

ER -