<?xml version="1.0" standalone="yes"?>
<Paper uid="W01-1407">
  <Title>Toward hierarchical models for statistical machine translation of inflected languages</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> The statistical approach to machine translation has become widely accepted in the last few years.</Paragraph>
    <Paragraph position="1"> It has been successfully applied to realistic tasks in various national and international research programs. However in many applications only small amounts of bilingual training data are available for the desired domain and language pair, and it is highly desirable to avoid at least parts of the costly data collection process.</Paragraph>
    <Paragraph position="2"> Some recent publications have dealt with the problem of translation with scarce resources.</Paragraph>
    <Paragraph position="3"> (Brown et al., 1994) describe the use of dictionaries. (Al-Onaizan et al., 2000) report on an experiment of Tetun-to-English translation by different groups, including one using statistical machine translation. They assume the absence of linguistic knowledge sources such as morphological analyzers and dictionaries. Nevertheless, they found that human mind is very well capable of deriving dependencies such as morphology, cognates, proper names, spelling variations etc., and that this capability was finally at the basis of the better results produced by humans compared to corpus based machine translation. The additional information results from complex reasoning and it is not directly accessible from the full word form representation of the data.</Paragraph>
    <Paragraph position="4"> In this paper, we take a different point of view: Even if full bilingual training data is scarce, monolingual knowledge sources like morphological analyzers and data for training the target language model as well as conventional dictionaries (one word and its translation per entry) may be available and of substantial usefulness for improving the performance of statistical translation systems. This is especially the case for highly inflected languages like German.</Paragraph>
    <Paragraph position="5"> We address the question of how to achieve a better exploitation of the resources for training the parameters for statistical machine translation by taking into account explicit knowledge about the languages under consideration. In our approach we introduce equivalence classes in order to ignore information not relevant to the translation process. We furthermore suggest the use of hierarchical lexicon models.</Paragraph>
    <Paragraph position="6"> The paper is organized as follows. After reviewing the statistical approach to machine translation, we first explain our motivation for examining the morphological characteristics of an inflected language like German. We then describe the chosen output representation after the analysis and present our approach for exploiting the information from morpho-syntactic analysis. Experimental results on the German-English Verbmobil task are reported.</Paragraph>
  </Section>
class="xml-element"></Paper>