Contrastive Entity Coreference and Disambiguation for Historical Texts

TL;DR

This paper introduces a large-scale dataset and contrastive bi-encoder models for improved entity coreference and disambiguation in historical texts, outperforming existing methods especially for out-of-knowledgebase individuals.

Contribution

It presents a novel large-scale training dataset, high-quality evaluation data, and trained models specifically designed for historical entity disambiguation and coreference resolution.

Findings

Models outperform existing disambiguation methods on historical benchmarks.

The approach effectively identifies out-of-knowledgebase individuals.

Models show competitive performance on modern entity disambiguation datasets.

Abstract

Massive-scale historical document collections are crucial for social science research. Despite increasing digitization, these documents typically lack unique cross-document identifiers for individuals mentioned within the texts, as well as individual identifiers from external knowledgebases like Wikipedia/Wikidata. Existing entity disambiguation methods often fall short in accuracy for historical documents, which are replete with individuals not remembered in contemporary knowledgebases. This study makes three key contributions to improve cross-document coreference resolution and disambiguation in historical texts: a massive-scale training dataset replete with hard negatives - that sources over 190 million entity pairs from Wikipedia contexts and disambiguation pages - high-quality evaluation data from hand-labeled historical newswire articles, and trained models evaluated on this historical benchmark. We contrastively train bi-encoder models for coreferencing and disambiguating individuals in historical texts, achieving accurate, scalable performance that identifies out-of-knowledgebase individuals. Our approach significantly surpasses other entity disambiguation models on our historical newswire benchmark. Our models also demonstrate competitive performance on modern entity disambiguation benchmarks, particularly certain news disambiguation datasets.