<?xml version="1.0" standalone="yes"?>
<Paper uid="W05-0612">
  <Title>An Expectation Maximization Approach to Pronoun Resolution</Title>
  <Section position="3" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> Coreference resolution is the process of determining which expressions in text refer to the same real-world entity. Pronoun resolution is the important yet challenging subset of coreference resolution where a system attempts to establish coreference between a pronominal anaphor, such as a third-person pronoun like he, she, it, or they, and a preceding noun phrase, called an antecedent. In the following example, a pronoun resolution system must determine the correct antecedent for the pronouns &amp;quot;his&amp;quot; and &amp;quot;he.&amp;quot; (1) When the president entered the arena with his family, he was serenaded by a mariachi band.</Paragraph>
    <Paragraph position="1"> Pronoun resolution has applications across many areas of Natural Language Processing, particularly in the field of information extraction. Resolving a pronoun to a noun phrase can provide a new interpretation of a given sentence, giving a Question Answering system, for example, more data to consider.</Paragraph>
    <Paragraph position="2"> Our approach is a synthesis of linguistic and statistical methods. For each pronoun, a list of antecedent candidates derived from the parsed corpus is presented to the Expectation Maximization (EM) learner. Special cases, such as pleonastic, reflexive and cataphoric pronouns are dealt with linguistically during list construction. This allows us to train on and resolve all third-person pronouns in a large Question Answering corpus. We learn lexicalized gender/number, language, and antecedent probability models. These models, tied to individual words, can not be learned with sufficient coverage from labeled data. Pronouns are resolved by choosing the most likely antecedent in the candidate list according to these distributions. The resulting resolution accuracy is comparable to supervised methods.</Paragraph>
    <Paragraph position="3"> We gain further performance improvement by initializing EM with a gender/number model derived from special cases in the training data. This model is shown to perform reliably on its own. We also demonstrate how the models learned through our unsupervised method can be used as features in a supervised pronoun resolution system.</Paragraph>
  </Section>
class="xml-element"></Paper>