Lexical Resources to Enrich English Malayalam Machine Translation

Lexical Resources to Enrich English Malayalam Machine Translation Sreelekha. S, Pushpak Bhattacharyya Dept. of Computer Science & Engg., IIT Bombay, Mumbai, India {sreelekha, pb}@cse.iitb.ac.in Abstract In this paper we present our work on the usage of lexical resources for the Machine Translation English and Malayalam. We describe a comparative performance between different Statistical Machine Translation (SMT) systems on top of phrase based SMT system as baseline. We explore different ways of utilizing lexical resources to improve the quality of English Malayalam statistical machine translation. In order to enrich the training corpus we have augmented the lexical resources in two ways (a) additional vocabulary and (b) inflected verbal forms. Lexical resources include IndoWordnet semantic relation set, lexical words and verb phrases etc. We have described case studies, evaluations and have given detailed error analysis for both Malayalam to English and English to Malayalam machine translation systems. We observed significant improvement in evaluations of translation quality. Lexical resources do help uplift performance when parallel corpora are scanty. Keywords: Lexical Resources, Statistical Machine Translation, English-Malayalam Machine Translation 1. Introduction Each Machine processing of Natural (Human) Languages has a long tradition, benefiting from decades of manual and semi-automatic analysis by linguists, sociologists, psychologists and computer scientists among others. Development of a full-fledged bilingual Machine Translation (MT) system for any two natural languages with limited electronic resources and tools is a challenging and demanding task. Since India is rich in linguistic divergence there are many morphologically rich languages quite different from English as well as from each other, there is a large requirement for machine translation between them. Development of efficient machine translation systems using appropriate methodologies and with limited resources is a challenging task. There are many ongoing attempts to develop MT systems for Indian languages (Antony, 2013; Kunchukuttan et al., 2014; Sreelekha et al., 2014; Sreelekha et al., 2015) using both rule based and statistical approaches. There were many attempts to improve the quality of Statistical MT systems such as; using Monolingually-Derived Paraphrases(Marton et al., 2009), Using Related Resource-Rich languages (Nakov and Ng, 2012) Considering the large amount of human effort and linguistic knowledge required for developing rule based systems, statistical MT systems became a better choice in terms of efficiency. Still the statistical systems fail to handle rich morphology. Consider the English sentence, He has been sent to the mosque for opening the door The English–Malayalam SMT system translated it as, Malayalam- അവൻ അയച്ചു mosque വാതിൽ തുറന്നു {avan ayachu mosque vathil thurannu} {He sent mosque door opened} {He sent mosque opened door} Here the system fails to translate the verb phrase “has been sent to” together and it translated a part of the phrase “sent” as “അയച്ചു”{ayachu}{sent}, which is wrong in the context. The same way another verb phrase “for opening the door” has been translated partly as “തുറന്നു” {thurannu} {opened}. Also, the system has deficiency in vocabulary and it couldn’t translate the English word “Mosque”. In these kinds of situations in order to learn various inflected forms and verb phrases, lexical …

Content truncated to 500 words