Parallel Corpora for bi-lingual English-Ethiopian Languages Statistical Machine Translation

Proceedings of the 27th International Conference on Computational Linguistics, pages 3102–3111 Santa Fe, New Mexico, USA, August 20-26, 2018. 3102 Parallel Corpora for bi-lingual English-Ethiopian Languages Statistical Machine Translation Solomon Teferra Abate1, Michael Melese Woldeyohannis1, Martha Yifiru Tachbelie1, Million Meshesha1, Solomon Atinafu1, Wondwossen Mulugeta1, Yaregal Assabie1, Hafte Abera1, Biniyam Ephrem1, Tewodros Abebe1, Wondimagegnhue Tsegaye2, Amanuel Lemma3, Tsegaye Andargie4, Seifedin Shifaw4 1Addis Ababa University, Addis Ababa, Ethiopia, 2Bahir Dar University, Bahir Dar, Ethiopia 3Aksum University, Axum, Ethiopia, 4Wolkite University, Wolkite, Ethiopia {solomon.teferra, michael.melese, martha.yifiru, million.meshesha, solomon.atnafu, wondwossen.mulugeta, yaregal.assabie, hafte.abera, binyam.ephrem, tewodros.abebe}@aau.edu.et, {wendeal, amanu.infosys, adtsegaye, seifedin28}@gmail.com Abstract In this paper, we describe an attempt towards the development of parallel corpora for En- glish and Ethiopian Languages, such as Amharic, Tigrigna, Afan-Oromo, Wolaytta and Ge’ez. The corpora are used for conducting a bi-directional statistical machine transla- tion experiments. The BLEU scores of the bi-directional Statistical Machine Translation (SMT) systems show a promising result. The morphological richness of the Ethiopian languages has a great impact on the performance of SMT specially when the targets are Ethiopian languages. Now we are working towards an optimal alignment for a bi- directional English-Ethiopian languages SMT. 1 Introduction The advancement of technology and the rise of the internet as a means of communication led to an ever increasing demand for Natural Language Processing (NLP) applications. NLP applications are useful in facilitating human-machine and human-human communications. One of the NLP applications which facilitates human-human communication is Machine Translation (MT). MT refers to a process by which computer software is used to translate a text or speech from one language to another (Koehn, 2009). In the presence of high volume digital text, the ideal aim of machine translation systems is to produce the best possible translation with minimal human intervention (Hutchins, 2005). The translation of natural language by machine becomes a reality, for technologically favored languages, in the late 20th century although it is dreamt in seventieth century (Hutchins, 1995). Various approaches to MT have been and are being used in the research community, that can broadly classified as rule-based and corpus based (Koehn, 2009). The rule based machine translation demands various kinds of linguistic resources such as morphological analyzer and synthesizer, syntactic parsers, semantic analyzers and so on. On the other hand, corpus based approaches (as the name implies) require parallel and monolingual corpora. Since corpus based approaches do not require deep linguistic analysis of the source and target languages, it is the preferred approach for under-resourced languages of the world, including Ethiopian languages. 1.1 Machine Translation For Ethiopian Languages Research in the development of MT has been conducted for technologically favored and econom- ically as well as politically important languages of the world since the 17th century. As a result, notable progress towards the development and use of MT systems has been made for these lan- guages. However, research in the area of MT for Ethiopian languages, which are under-resourced as well as economically and technologically disadvantaged, has started very recently. Most of the This work is licenced …

Content truncated to 500 words