<?xml version="1.0" standalone="yes"?>
<Paper uid="P06-1060">
  <Title>Factorizing Complex Models: A Case Study in Mention Detection</Title>
  <Section position="6" start_page="479" end_page="479" type="concl">
    <SectionTitle>
5 Conclusion
</SectionTitle>
    <Paragraph position="0"> As natural language processing becomes more sophisticated and powerful, we start focus our attention on more and more properties associated with the objects we are seeking, as they allow for a deeper and more complex representation of the real world. With this focus comes the question of how this goal should be accomplished - either detect all properties at once, one at a time through a pipeline, or a hybrid model. This paper presents three methods through which multi-label sequence classification can be achieved, and evaluates and contrasts them on the Automatic Content Extraction task. On the ACE mention detection task, the cascade model which predicts first the mention boundaries and entity types, followed by mention  typeandentitysubtypeoutperformsthesimpleallin-one model in most cases, and the joint model in a few cases.</Paragraph>
    <Paragraph position="1"> Among the proposed models, the cascade approach has the definite advantage that it can easily and productively incorporate additional partially-labeled data. We also presented a novel modification of the joint system training that allows for the direct incorporation of additional data, which increased the system performance significantly. The all-in-one model can only incorporate additional data in an indirect way, resulting in little to no overall improvement.</Paragraph>
    <Paragraph position="2"> Finally, the performance obtained by the cascademodelisverycompetitive: whenpairedwitha coreference module, it ranked very well in the &amp;quot;Entity Detection and Tracking&amp;quot; task in the ACE'04 evaluation.</Paragraph>
  </Section>
class="xml-element"></Paper>