<?xml version="1.0" standalone="yes"?>
<Paper uid="A00-3004">
  <Title>A Weighted Robust Parsing Approach to Semantic Annotation</Title>
  <Section position="6" start_page="21" end_page="21" type="concl">
    <SectionTitle>
3 Conclusions and future works
</SectionTitle>
    <Paragraph position="0"> In this paper we summarized a proposal for a framework for designing grammar-based automated annotation applications. Starting with a case study and following an approach which combines the notions of fuzziness and robustness in sentence parsing, we showed how to build practical domain-dependent rules which can be applied whenever it is possible to superimpose a sentence-level semantic structure to a text without relying on a previous deep syntactical analysis. Even if the query generation problem may not seem a critical application one should bear in mind that the sentence processing must be done on-line.</Paragraph>
    <Paragraph position="1"> As we have previously seen, the cue-words used as semantic markers are domain-dependent. Even their relevance disposal and their weight within the rules depends on their linguistic usage. Therefore, a complete automatic annotation system based on the approach proposed in this article seems to be adequate to give precise results. However, a semi-automatic system could satisfy our needs. This system should be based on the following techniques to achieve a high level of performance: 1. For each annotation, the system offers a list of propositions based on standard grammars as well as on external knowledge (ontologies, knowledge bases ...) 2. According to the grammar initially proposed, the user may change the annotation according to his needs. These modifications are held within the system to change the grammar rules as well as their weights. This makes the system interactive and enhanced by a learning phase.</Paragraph>
    <Paragraph position="2"> 3. We could imagine that rule design process can be partially automated and we intend to pursue some research on developing methods for both assisted rule design and corpus based rule induction. null</Paragraph>
  </Section>
class="xml-element"></Paper>