Multi-Attribute Relation Extraction (MARE) -- Simplifying the Application of Relation Extraction

TL;DR

This paper introduces MARE, a simplified, assumption-less approach to multi-attribute relation extraction that reduces annotation complexity and improves extraction performance over existing methods.

Contribution

The paper proposes a novel assumption-less formulation for multi-attribute relation extraction and two models that simplify annotation requirements and enhance extraction accuracy.

Findings

MARE outperforms state-of-the-art event and binary relation extraction methods.

Simplified annotation process facilitates easier application in real-world scenarios.

Models show improved extraction of general multi-attribute relations.

Abstract

Natural language understanding's relation extraction makes innovative and encouraging novel business concepts possible and facilitates new digitilized decision-making processes. Current approaches allow the extraction of relations with a fixed number of entities as attributes. Extracting relations with an arbitrary amount of attributes requires complex systems and costly relation-trigger annotations to assist these systems. We introduce multi-attribute relation extraction (MARE) as an assumption-less problem formulation with two approaches, facilitating an explicit mapping from business use cases to the data annotations. Avoiding elaborated annotation constraints simplifies the application of relation extraction approaches. The evaluation compares our models to current state-of-the-art event extraction and binary relation extraction methods. Our approaches show improvement compared to these on the extraction of general multi-attribute relations.