NLP Infrastructure for the Lithuanian Language

NLP Infrastructure for the Lithuanian Language Daiva Vitkutė-Adžgauskienė, Andrius Utka, Darius Amilevičius, Tomas Krilavičius Vytautas Magnus University, K. Donelaičio g. 58, 44248 Kaunas, Lithuania E-mail: {d.vitkute-adzgauskiene, d.amilevicius, t.krilavicius}@if.vdu.lt, a.utka@hmf.vdu.lt Abstract The Information System for Syntactic and Semantic Analysis of the Lithuanian language (lith. Lietuvių kalbos sintaksinės ir semantinės analizės informacinė sistema, LKSSAIS) is the first infrastructure for the Lithuanian language combining Lithuanian language tools and resources for diverse linguistic research and applications tasks. It provides access to the basic as well as advanced natural language processing tools and resources, including tools for corpus creation and management, text preprocessing and annotation, ontology building, named entity recognition, morphosyntactic and semantic analysis, sentiment analysis, etc. It is an important platform for researchers and developers in the field of natural language technology. Keywords: language technology infrastructure, Lithuanian language 1. Introduction The paper presents LKSSAIS1, the first infrastructure that provides online access to open-source tools for managing, processing, annotating, and analyzing Lithuanian language texts. The system was designed according to the best practices in the development of language technology infrastructures, keeping in mind the possibility of integration with the European CLARIN ERIC Infrastructure 2 . It attempts to address the needs of researchers working with the Lithuanian language, taking into account continuously growing amount of data and growing need for natural language processing applications in scientific, public, and business fields. It also provides online free access to public services for syntactic and semantic analysis of Lithuanian language texts. Language tools and resources today are of paramount importance for research and teaching, as well as for different industrial applications. Two main factors leading to the development of Lithuanian language technology infrastructure were: 1) Lithuanian language has been considered to be extremely under-resourced on the NLP side and had a very weak support according to META-NET survey (Vaišnienė and Zabarskaitė 2012); 2) Lithuanian language is very rich in inflections and diacritics, therefore reuse or simple adaptation of widely-used English language technology tools and specifications was unreasonable. Although prior to 2013, there had been a number of activities for developing language tools, resources and services for the Lithuanian language (Marcinkevičienė and Vitkutė-Adžgauskienė 2010), these activities were not well coordinated, separate tools and corpora were created using different standards, without well documented APIs and without open-access interfaces. Unsurprisingly the quality of most of NLP tools for Lithuanian was far away from the state of the art. In order to improve such situation, a project for the 1 http://semantika.lt 2 http://clarin.eu development of the Information System for Syntactic and Semantic Analysis of the Lithuanian language was launched in the middle of 2012. The project was the collaboration of the Vytautas Magnus University (coordinator) and Kaunas Technology University, and its first results were presented in May, 2015. The main result of the project is an infrastructure combining Lithuanian language tools and resources for diverse language research and applications tasks, thus representing the first national NLP infrastructure for the Lithuanian language. It should be noted that most tools and services for the …

Content truncated to 500 words