<?xml version="1.0" standalone="yes"?>
<Paper uid="W06-3101">
  <Title>Morpho-syntactic Information for Automatic Error Analysis of Statistical Machine Translation Output Maja Popovi'cstar</Title>
  <Section position="3" start_page="0" end_page="1" type="intro">
    <SectionTitle>
2 Related Work
</SectionTitle>
    <Paragraph position="0"> There is a number of publications dealing with various automatic evaluation measures for machine translation output, some of them proposing new measures, some proposing improvements and extensions of the existing ones (Doddington, 2002; Papineni et al., 2002; Babych and Hartley, 2004; Matusov et al., 2005). Semi-automatic evaluation measures have been also investigated, for example in (Niessen et al., 2000). An automatic metric which uses base forms and synonyms of the words in order to correlate better to human judgements has been  proposed in (Banerjee and Lavie, 2005). However, error analysis is still a rather unexplored area. A framework for human error analysis and error classification has been proposed in (Vilar et al., 2006) and a detailed analysis of the obtained results has been carried out. Automatic methods for error analysis to our knowledge have not been studied yet.</Paragraph>
    <Paragraph position="1"> Many publications propose the use of morpho-syntactic information for improving the performance of a statistical machine translation system. Various methods for treating morphological and syntactical differences between German and English are investigated in (Niessen and Ney, 2000; Niessen and Ney, 2001a; Niessen and Ney, 2001b). Morphological analysis has been used for improving Arabic-English translation (Lee, 2004), for Serbian-English translation (Popovi'c et al., 2005) as well as for Czech-English translation (Goldwater and Mc-Closky, 2005). Inflectional morphology of Spanish verbs is dealt with in (Popovi'c and Ney, 2004; de Gispert et al., 2005). To the best of our knowledge, the use of morpho-syntactic information for error analysis of translation output has not been investigated so far.</Paragraph>
  </Section>
class="xml-element"></Paper>