<?xml version="1.0" standalone="yes"?>
<Paper uid="P03-1012">
  <Title>A Probability Model to Improve Word Alignment</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> Word alignments were first introduced as an intermediate result of statistical machine translation systems (Brown et al., 1993). Since their introduction, many researchers have become interested in word alignments as a knowledge source. For example, alignments can be used to learn translation lexicons (Melamed, 1996), transfer rules (Carbonell et al., 2002; Menezes and Richardson, 2001), and classifiers to find safe sentence segmentation points (Berger et al., 1996).</Paragraph>
    <Paragraph position="1"> In addition to the IBM models, researchers have proposed a number of alternative alignment methods. These methods often involve using a statistic such as ph2 (Gale and Church, 1991) or the log likelihood ratio (Dunning, 1993) to create a score to measure the strength of correlation between source and target words. Such measures can then be used to guide a constrained search to produce word alignments (Melamed, 2000).</Paragraph>
    <Paragraph position="2"> It has been shown that once a baseline alignment has been created, one can improve results by using a refined scoring metric that is based on the alignment. For example Melamed uses competitive linking along with an explicit noise model in (Melamed, 2000) to produce a new scoring metric, which in turn creates better alignments.</Paragraph>
    <Paragraph position="3"> In this paper, we present a simple, flexible, statistical model that is designed to capture the information present in a baseline alignment. This model allows us to compute the probability of an alignment for a given sentence pair. It also allows for the easy incorporation of context-specific knowledge into alignment probabilities.</Paragraph>
    <Paragraph position="4"> A critical reader may pose the question, &amp;quot;Why invent a new statistical model for this purpose, when existing, proven models are available to train on a given word alignment?&amp;quot; We will demonstrate experimentally that, for the purposes of refinement, our model achieves better results than a comparable existing alternative.</Paragraph>
    <Paragraph position="5"> We will first present this model in its most general form. Next, we describe an alignment algorithm that integrates this model with linguistic constraints in order to produce high quality word alignments. We will follow with our experimental results and discussion. We will close with a look at how our work relates to other similar systems and a discussion of possible future directions.</Paragraph>
  </Section>
class="xml-element"></Paper>