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

Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 1547–1557 Copenhagen, Denmark, September 7–11, 2017. c⃝2017 Association for Computational Linguistics Identifying civilians killed by police with distantly supervised entity-event extraction Katherine A. Keith, Abram Handler, Michael Pinkham, Cara Magliozzi, Joshua McDuffie, and Brendan O’Connor College of Information and Computer Sciences University of Massachusetts Amherst kkeith@cs.umass.edu, brenocon@cs.umass.edu http://slanglab.cs.umass.edu/ 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 neu- ral network classifiers. Our model out- performs two off-the-shelf event extractor systems, and it can suggest candidate vic- tim names in some cases faster than one of the major manually-collected police fatal- ity databases. Appendix, software, and data are avail- able online at: http://slanglab.cs.umass. edu/PoliceKillingsExtraction/ 1 Introduction The United States government does not keep sys- tematic records of when police kill civilians, de- spite a clear need for this information to serve the public interest and support social scientific anal- ysis. Federal records rely on incomplete cooper- ation from local police departments, and human rights statisticians assess that they fail to document thousands of fatalities (Lum and Ball, 2015). News articles have emerged as a valuable al- ternative data source. Organizations including The Guardian, The Washington Post, Mapping Po- lice Violence, and Fatal Encounters have started to build such databases of U.S. police killings by manually reading millions of news articles1 1Fatal Encounters director D. Brian Burghart estimates he and colleagues have read 2 million news headlines and ledes to assemble its fatality records that date back to January, 2000 (pers. comm.); we find FE to be the most comprehensive pub- licly available database. Text Person killed by police? Alton Sterling was killed by police. True Officers shot and killed Philando Castile. True Officer Andrew Hanson was shot. False Police report Megan Short was fatally shot in apparent murder-suicide. False Table 1: Toy examples (with entities in bold) illus- trating the problem of extracting from text names of persons who have been killed by police. and extracting victim names and event details. This approach was recently validated by a Bu- reau of Justice Statistics study (Banks et al., Dec. 2016) which augmented traditional police- maintained records with media reports, finding twice as many deaths compared to past govern- ment analyses. This suggests textual news data has enormous, real value, though manual news analy- sis remains extremely laborious. We propose to help automate this process by ex- tracting the names of persons killed by police from event descriptions in news articles (Table 1). This can be formulated as either of two cross-document entity-event extraction tasks: 1. Populating an entity-event database: From a corpus of news articles D(test) over timespan T, extract the names of persons killed by po- lice during that same timespan (E(pred)). …

Content truncated to 500 words