Abstract:Relation Extraction (RE), the task of detecting and characterizing semantic relations between entities in text, has gained much importance in the last two decades, mainly in the biomedical domain. Many papers have been published on Relation Extraction using supervised machine learning techniques. Most of these techniques rely on statistical methods, such as feature-based and tree-kernels-based methods. Such statistical learning techniques are usually based on a propositional hypothesis space for representing examples, i.e., they employ an attribute-value representation of features. This kind of representation has some drawbacks, particularly in the extraction of complex relations which demand more contextual information about the involving instances, i.e., it is not able to effectively capture structural information from parse trees without loss of information. In this work, we present OntoILPER, a logic-based relational learning approach to Relation Extraction that uses Inductive Logic Programming for generating extraction models in the form of symbolic extraction rules. OntoILPER takes profit of a rich relational representation of examples, which can alleviate the aforementioned drawbacks. The proposed relational approach seems to be more suitable for Relation Extraction than statistical ones for several reasons that we argue. Moreover, OntoILPER uses a domain ontology that guides the background knowledge generation process and is used for storing the extracted relation instances. The induced extraction rules were evaluated on three protein-protein interaction datasets from the biomedical domain. The performance of OntoILPER extraction models was compared with other state-of-the-art RE systems. The encouraging results seem to demonstrate the effectiveness of the proposed solution.

Domain Ontology Concept and Relation Extraction Using Log-Likelihood Ratio

Study on Context-Based Domain Ontology Concept Extraction and Relation Extraction

Ontology Matching Based On Probabilistic Description Logic

Domain relation learning from events in log ontologies based on hierarchy association rules

Towards Context-Sensitive Domain Ontology Extraction

A Constructivist Ontology Relation Learning Method

Domain Ontology Learning Enhanced by Optimized Relation Instance in DBpedia.

Learning Logic Rules for Document-level Relation Extraction

Logic-guided Semantic Representation Learning for Zero-Shot Relation Classification.

Learning Relation Instances for Chinese Domain Ontology from the Web

A logic-based relational learning approach to relation extraction: The OntoILPER system

A Practical System of Domain Ontology Learning Using the Web for Chinese

Learning Ontology From Relational Database

An Integrated Approach for Concept Learning and Relation Extraction

Automatic Domain-Ontology Relation Extraction from Semi-structured Texts

A Domain-Independent Ontology Learning Method Based on Transfer Learning

A Term Extraction Approach Based on Modified Log-likelihood Ratio

Chinese domain ontology learning based on instance extraction and classification

Relation extraction for ontology extension using integrated evidences

Do LLMs Really Adapt to Domains? An Ontology Learning Perspective

Design and Implementation of Domain Ontology for Information Extraction