<?xml version="1.0" standalone="yes"?>
<Paper uid="W05-0304">
  <Title>Parallel Entity and Treebank Annotation</Title>
  <Section position="2" start_page="0" end_page="21" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> A great deal of annotation effort for many different corpora has been devoted to annotation for entities and syntactic structure (treebanks). However, previous efforts at treebanking have largely been independent of the constituency of entities, and previous efforts at entity annotation have likewise been independent of corresponding layers of syntactic structure. We describe here a corpus being developed for biomedical information extraction with levels of both entity annotation and treebank annotation, with a goal that entities can be mapped to constituents in the treebank.</Paragraph>
    <Paragraph position="1"> We are collaborating with researchers in the Division of Oncology at The Children's Hospital of Philadelphia, for the purpose of automatically mining the corpus of cancer literature for those associations that link specified variations in individual genes with known malignancies. In particular, we are interested in extracting three entities (Gene, Variation event, and Malignancy) in the following relationship: Gene X with genomic Variation event Y is correlated with Malignancy Z. For example, WT1 is deleted in Wilms Tumor #5. In addition, Variation events are themselves relations, consisting of entities representing different aspects of a Variation event.</Paragraph>
    <Paragraph position="2"> Mapping entities to treebank constituents is a desirable goal since the entities can then be viewed as semantic types associated with syntactic constituents, and we expect that automated analyses of these related levels will interact in a mutually reinforcing and beneficial way for development of statistical taggers.</Paragraph>
    <Paragraph position="3"> In this paper we describe aspects of the entity and treebank annotation that allow this mapping to be largely successful. Potentially large entities that would otherwise cut across syntactic constituents are decomposed into components of a relation. While this is worthwhile by itself on conceptual grounds for entity definition, and was in fact not done for reasons of mapping to syntactic constituents, it makes such a mapping easier. The tree-bank annotation has been modified from the Penn Treebank guidelines in various ways, such as greater structure for prenominal modifiers. Again, while this would have been done regardless of the mapping of entities, it does make such a mapping more successful.</Paragraph>
    <Paragraph position="4"> Previous work on integrating syntactic structure with entity information, as well as relation infor- null mation, is described in (Miller et al., 2000). Our work is in much the same spirit, although we do not integrate relation annotation into the syntactic trees. PubMed abstracts are quite different from the newswire sources used in that earlier work, with several consequences discussed throughout, such as the use of discontinuous entities.</Paragraph>
    <Paragraph position="5"> Section 2 discusses some of the main issues around the development of the guidelines for entity annotation, and Section 3 discusses some of the changes that have been made for the treebank guidelines. Section 4 describes the annotation workflow and the resulting merged representation. Section 5 evaluates the mapping between entities and constituents, and Section 6 is the conclusion.</Paragraph>
  </Section>
class="xml-element"></Paper>