<?xml version="1.0" standalone="yes"?>
<Paper uid="W03-1902">
  <Title>PIPCA - Unisinos</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> We have been dealing with corpus based studies since 1997 (Renata Vieira and Simone Teufel, 1997; Poesio et al., 1997). Our focus has been the study of coreference. In the study of coreference we have dealt with annotation experiments (manual and automatic) and their respective annotation schemes. To work on coreference we used information from syntactic annotated corpus, the Penn Treebank. Our results (annotated corpus with coreference links and classification of coreference status) were Prolog encoded. When we first adapted our tool for Portuguese (Rossi et al., 2001) we dealt with other tools and annotation formats. The resources built on these previous works were difficult to share due to their particular information encoding.</Paragraph>
    <Paragraph position="1"> Our current work in the COMMOn-REFs project (A computational model for processing referring ex- null Research Grant CNPq- Brazil.</Paragraph>
    <Paragraph position="2"> are using MMAX, a tool for multimodal annotation in XML (Muller and Strube, 2001), for manual annotation of coreference, and we are developing a tool for automatic coreference resolution. Our tool deals with XML encoding provided by MMAX and syntactic information for Portuguese and French encoded in XML. In order to be able to share the resources being built, we are relating our model with proposed standards.</Paragraph>
    <Paragraph position="3"> In Section 2 we present previous annotation formats that we dealt with. In Section 3 we give an overview of the work in COMMOn-REFs. Section 4 relates our current model with the standards recently proposed (Ide and Romary, 2002; Ide and Romary, 2003; Ide and Romary, 2001). Section 5 describes our tool for coreference resolution. A discussion on the problems we face with our annotation model is presented in Section 6.</Paragraph>
  </Section>
class="xml-element"></Paper>