Identifying civilians killed by police with distantly supervised entity-event extraction

TL;DR

This paper introduces a new NLP task to identify police victims from news articles, using a novel corpus and a model that outperforms existing systems in extracting victim names efficiently.

Contribution

The paper presents a publicly available police fatality corpus and a novel EM-based distant supervision model for extracting victim names from news texts.

Findings

Model outperforms existing event extractors.

Can suggest victim candidates faster than some manual databases.

Provides a new dataset for police fatality detection.

Abstract

We propose a new, socially-impactful task for natural language processing: from a news corpus, extract names of persons who have been killed by police. We present a newly collected police fatality corpus, which we release publicly, and present a model to solve this problem that uses EM-based distant supervision with logistic regression and convolutional neural network classifiers. Our model outperforms two off-the-shelf event extractor systems, and it can suggest candidate victim names in some cases faster than one of the major manually-collected police fatality databases.