Annotation schemes in North Sámi dependency parsing

The 3rd International Workshop for Computational Linguistics of Uralic Languages, pages 66–75, St. Petersburg, Russia, 23–24 January 2017. c⃝2017 Association for Computational Linguistics Annotation schemes in North Sámi dependency parsing Mariya Sheyanova Fundamental and Applied Linguistics Higher School of Economics Moscow, Russia masha.shejanova@gmail.com Francis M. Tyers Giela ja kultuvrra instituhtta UiT Norgga árktalaš universitehta N-9018 Romsa, Norway francis.tyers@uit.no Abstract In this paper we describe a comparison of two annotation schemes for de- pendency parsing of North Sámi, a Finno-Ugric language spoken in the north of Scandinavia and Finland. The two annotation schemes are the Giellatekno (GT) scheme which has been used in research and applications for the Sámi languages and Universal Dependencies (UD) which is a cross-lingual scheme aiming to unify annotation stations across languages. We show that we are able to deterministi- cally convert from the Giellatekno scheme to the Universal Dependencies scheme without a loss of parsing performance. While we do not claim that either scheme is a priori a more adequate model of North Sámi syntax, we do argue that the choice of annotation scheme is dependent on the intended application. This work is licensed under a Creative Commons Attribution–NoDerivatives 4.0 International Licence. Licence details: http://creativecommons.org/licenses/by-nd/4.0/ 1 66 1 Introduction Dependency parsing is an important step in many applications of natural language processing, such as information extraction, machine translation, interactive language learning and corpus search interfaces. There are a number of approaches to depen- dency parsing. These include rule-based approaches such as [1] for German and [2] for Portuguese, and statistical approaches, such as transition-based parsing, exem- plified by MaltParser [3]. In rule-based approaches the knowledge source is a set of rules (such as constraints), while in statistical approaches the knowledge source is a collection of parsed sentences called treebank. Both rule-based and statistical approaches have in common that the parses they output conform to a given annotation scheme, that is a set of rules which define what linguistic structure should be applied to given constructions. Annotation schemes can vary substantially according to how they encode an analysis of linguistic struc- ture. For example, one scheme may decide that the auxiliary verb is the head of an auxiliary—main verb construction because of subject agreement, while another may decide that the main verb is the head because of case government of arguments. The choice of representation can depend heavily on application. For certain ap- plications, such as grammar checking, a more morphosyntactic scheme may be ap- propriate, while for others, such as machine translation, a more syntacto-semantic scheme may be more appropriate. In this paper, we describe the conversion of the annotation schemes output by the Giellatekno parser to a corresponding UD-compliant scheme. We also provide com- parative statistics on the composition of the corpus used in the experiments before and after the conversion. Afterwards, we report a comparison of parsing results using the two annotation schemes. The paper is structured as follows: Section 2 describes the corpus and source annotation scheme used in this paper; section 3 …

Content truncated to 500 words