<?xml version="1.0" standalone="yes"?>
<Paper uid="H05-1004">
  <Title>Proceedings of Human Language Technology Conference and Conference on Empirical Methods in Natural Language Processing (HLT/EMNLP), pages 25-32, Vancouver, October 2005. c(c)2005 Association for Computational Linguistics On Coreference Resolution Performance Metrics</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> A working definition of coreference resolution is partitioning the noun phrases we are interested in into equivalence classes, each of which refers to a physical entity.</Paragraph>
    <Paragraph position="1"> We adopt the terminologies used in the Automatic Content Extraction (ACE) task (NIST, 2003a) and call each individual phrase a mention and equivalence class an entity. For example, in the following text segment,</Paragraph>
  </Section>
class="xml-element"></Paper>