<?xml version="1.0" standalone="yes"?>
<Paper uid="N06-1037">
  <Title>Exploring Syntactic Features for Relation Extraction using a Convolution Tree Kernel</Title>
  <Section position="9" start_page="294" end_page="294" type="concl">
    <SectionTitle>
5 Conclusion and Future Work
</SectionTitle>
    <Paragraph position="0"> In this paper, we explore the syntactic features using convolution tree kernels for relation extraction.</Paragraph>
    <Paragraph position="1"> We conclude that: 1) the relations between entities can be well represented by parse trees with carefully calibrating effective portions of parse trees; 2) the syntactic features embedded in a parse tree are particularly effective for relation extraction; 3) the convolution tree kernel can effectively capture the syntactic features for relation extraction.</Paragraph>
    <Paragraph position="2"> The most immediate extension of our work is to improve the accuracy of relation detection. We may adopt a two-step method (Culotta and Sorensen, 2004) to separately model the relation detection and characterization issues. We may integrate more features (such as head words or WordNet semantics) into nodes of parse trees. We can also benefit from the learning algorithm to study how to solve the data imbalance and sparseness issues from the learning algorithm viewpoint. In the future, we would like to test our algorithm on the other version of the ACE corpus and to develop fast algorithm (Vishwanathan and Smola, 2002) to speed up the training and testing process of convolution kernels.</Paragraph>
    <Paragraph position="3"> Acknowledgements: We would like to thank Dr.</Paragraph>
    <Paragraph position="4"> Alessandro Moschitti for his great help in using his Tree Kernel Toolkits and fine-tuning the system.</Paragraph>
    <Paragraph position="5"> We also would like to thank the three anonymous reviewers for their invaluable suggestions.</Paragraph>
  </Section>
class="xml-element"></Paper>