<?xml version="1.0" standalone="yes"?>
<Paper uid="P05-1057">
  <Title>Log-linear Models for Word Alignment</Title>
  <Section position="4" start_page="0" end_page="459" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> Word alignment, which can be defined as an object for indicating the corresponding words in a parallel text, was first introduced as an intermediate result of statistical translation models (Brown et al., 1993). In statistical machine translation, word alignment plays a crucial role as word-aligned corpora have been found to be an excellent source of translation-related knowledge.</Paragraph>
    <Paragraph position="1"> Various methods have been proposed for finding word alignments between parallel texts. There are generally two categories of alignment approaches: statistical approaches and heuristic approaches.</Paragraph>
    <Paragraph position="2"> Statistical approaches, which depend on a set of unknown parameters that are learned from training data, try to describe the relationship between a bilingual sentence pair (Brown et al., 1993; Vogel and Ney, 1996). Heuristic approaches obtain word alignments by using various similarity functions between the types of the two languages (Smadja et al., 1996; Ker and Chang, 1997; Melamed, 2000). The central distinction between statistical and heuristic approaches is that statistical approaches are based on well-founded probabilistic models while heuristic ones are not. Studies reveal that statistical alignment models outperform the simple Dice coefficient (Och and Ney, 2003).</Paragraph>
    <Paragraph position="3"> Finding word alignments between parallel texts, however, is still far from a trivial work due to the diversity of natural languages. For example, the alignment of words within idiomatic expressions, free translations, and missing content or function words is problematic. When two languages widely differ in word order, finding word alignments is especially hard. Therefore, it is necessary to incorporate all useful linguistic information to alleviate these problems. null Tiedemann (2003) introduced a word alignment approach based on combination of association clues.</Paragraph>
    <Paragraph position="4"> Clues combination is done by disjunction of single clues, which are defined as probabilities of associations. The crucial assumption of clue combination that clues are independent of each other, however, is not always true. Och and Ney (2003) proposed Model 6, a log-linear combination of IBM translation models and HMM model. Although Model 6 yields better results than naive IBM models, it fails to include dependencies other than IBM models and HMM model. Cherry and Lin (2003) developed a  statistical model to find word alignments, which allow easy integration of context-specific features.</Paragraph>
    <Paragraph position="5"> Log-linear models, which are very suitable to incorporate additional dependencies, have been successfully applied to statistical machine translation (Och and Ney, 2002). In this paper, we present a framework for word alignment based on log-linear models, allowing statistical models to be easily extended by incorporating additional syntactic dependencies. We use IBM Model 3 alignment probabilities, POS correspondence, and bilingual dictionary coverage as features. Our experiments show that log-linear models significantly outperform IBM translation models.</Paragraph>
    <Paragraph position="6"> We begin by describing log-linear models for word alignment. The design of feature functions is discussed then. Next, we present the training method and the search algorithm for log-linear models. We will follow with our experimental results and conclusion and close with a discussion of possible future directions.</Paragraph>
  </Section>
class="xml-element"></Paper>