Diachronic Analysis of Entities by Exploiting Wikipedia Page revisions

Proceedings of Recent Advances in Natural Language Processing, pages 84–91, Varna, Bulgaria, Sep 2–4, 2019. https://doi.org/10.26615/978-954-452-056-4_011 84 Diachronic Analysis of Entities by Exploiting Wikipedia Page revisions Pierpaolo Basile University of Bari Aldo Moro Dept. of Computer Science Bari, Italy pierpaolo.basile@uniba.it Annalina Caputo ADAPT Centre Trinity College Dublin Dublin, Ireland annalina.caputo@adaptcentre.ie Seamus Lawless ADAPT Centre Trinity College Dublin Dublin, Ireland seamus.lawless@adaptcentre.ie Giovanni Semeraro University of Bari Aldo Moro Dept. of Computer Science Bari, Italy giovanni.semeraro@uniba.it Abstract In the last few years, the increasing avail- ability of large corpora spanning several time periods has opened new opportunities for the diachronic analysis of language. This type of analysis can bring to the light not only linguistic phenomena related to the shift of word meanings over time, but it can also be used to study the impact that societal and cultural trends have on this language change. This paper intro- duces a new resource for performing the diachronic analysis of named entities built upon Wikipedia page revisions. This re- source enables the analysis over time of changes in the relations between entities (concepts), surface forms (words), and the contexts surrounding entities and surface forms, by analysing the whole history of Wikipedia internal links. We provide some useful use cases that prove the impact of this resource on diachronic studies and de- lineate some possible future usage. 1 Introduction The availability of large corpora spanning differ- ent time periods has encouraged researchers to analyse language from a diachronic perspective. Language is dynamic and detecting significant lin- guistic shifts in the meaning and usage of words is a crucial task for both social and cultural stud- ies and for Natural Language Processing applica- tions. Recent work focusing on the automatic de- tection of the semantic shift of words has adopted diachronic (or dynamic) word embeddings (Kim et al., 2014; Hamilton et al., 2016b; Kulkarni et al., 2015). This type of work represents words as vectors in a semantic space where the proxim- ity between word vectors indicate the existence of a semantic relationship between the terms in- volved. The diachronic analysis is then performed by building a different semantic space for each pe- riod of investigation and aligning vectors belong- ing to different spaces in order to make them com- parable. The variations in the similarity between the word vectors in two different spaces marks possible changes in the context of appearance of that word. This is used as a proxy indicator of change, either cultural, social or semantic, asso- ciated with the occurrence of that specific word. This kind of work has generated a variety of re- sources for the diachronic analysis of word mean- ings, covering different time periods, languages, and genres. While the broader area of automatic detection of semantic shift of words is gaining momentum, only little effort has focused on the more specific problem of analysing the semantic shift of named entities. This problem has a huge impact on the correct identification of entities in context, with repercussions on many natural …

Content truncated to 500 words