<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-0805">
  <Title>The Italian Lexical Sample Task at SENSEVAL-3</Title>
  <Section position="5" start_page="2" end_page="2" type="concl">
    <SectionTitle>
4 Conclusions
</SectionTitle>
    <Paragraph position="0"> The results of the six systems participating in the evaluation exercise showed some improvements compared to the average performance at SENSEVAL-2, though data sets and sense repositories were considerably different.</Paragraph>
    <Paragraph position="1"> We are pleased with the successful outcome of the experiments in terms of participation, although regrettably no system exploited the unlabeled training set, which was intended to offer a less time-consuming resource. On the other hand, the labeled instances that have been collected represent a useful and reusable benchmark.</Paragraph>
    <Paragraph position="2"> As a final remark we think it could be interesting to consider the actual distribution of word senses in Italian corpora in collecting the examples for the next campaign.</Paragraph>
  </Section>
class="xml-element"></Paper>