Axolotl: a Web Accessible Parallel Corpus for Spanish-Nahuatl

Axolotl: a Web Accessible Parallel Corpus for Spanish-Nahuatl Ximena Gutierrez-Vasques, Gerardo Sierra, Isaac Hernandez GIL IINGEN Instituto de Ingenier´ıa, UNAM, Mexico City, Mexico xim@unam.mx, {gsierram,ihernandezpo}@iingen.unam.mx Abstract This paper describes the project called Axolotl which comprises a Spanish-Nahuatl parallel corpus and its search interface. Spanish and Nahuatl are distant languages spoken in the same country. Due to the scarcity of digital resources, we describe the several problems that arose when compiling this corpus: most of our sources were non-digital books, we faced errors when digitizing the sources and there were difficulties in the sentence alignment process, just to mention some. The documents of the parallel corpus are not homogeneous, they were extracted from different sources, there is dialectal, diachronical, and orthographical variation. Additionally, we present a web search interface that allows to make queries through the whole parallel corpus, the system is capable to retrieve the parallel fragments that contain a word or phrase searched by a user in any of the languages. To our knowledge, this is the first Spanish-Nahuatl public available digital parallel corpus. We think that this resource can be useful to develop language technologies and linguistic studies for this language pair. Keywords: Parallel corpus, Low-resource languages, Search tool 1. Introduction Parallel corpora are a rich linguistic resource which com- prises bodies of text in parallel translation, i.e., a set of texts in different languages which are translations from each other. This type of corpus is one of the most valu- able resources in the development of several Natural Lan- guage Processing (NLP) applications. For instance, parallel corpora provides indispensable training data for the statis- tical machine translation systems (Brown et al., 1993) and it is also useful for some other applications like multilin- gual text retrieval and automatic bilingual lexical acquisi- tion (Widdows et al., 2002; Guinovart, 2012). Parallel texts are also a useful resource for aiding hu- man translators. For instance, computer assisted transla- tion tools or parallel corpus search interfaces, allow users to search for words and expressions in bilingual texts and see how other people translated an expression and in which context a certain translation is used (Volk et al., 2014). Additionally, parallel corpora can be used in the linguis- tics field, since it is helpful for performing contrastive and translation studies (Johansson, 2007). The most common sources for gathering large amounts of parallel data include specialized domain texts such as par- liamentary proceedings, religious texts and software manu- als. Additionally, the World Wide Web represents a good source for finding large-size and balanced parallel text (Resnik and Smith, 2003). On the Web it is possible to find bilingual or multilingual websites which can be useful to extract readily available parallel text in several domains and language pairs. There are several examples of websites that offer their contents in several languages, for instance: in- ternational institutions, universities, touristic services, etc. However, only for few language pairs there exist large amounts of readily available parallel data. We face a scarcity problem specially when one or …

Content truncated to 500 words