<?xml version="1.0" standalone="yes"?>
<Paper uid="C00-2166">
  <Title>Robust Semantic Construction</Title>
  <Section position="7" start_page="1104" end_page="1104" type="concl">
    <SectionTitle>
7 Summary
</SectionTitle>
    <Paragraph position="0"> The paper has presented a module d capable o:\[' tlandling inlmt from tlat, analysis lnetho(ls and transforming them into full-fledged semantic representations. The module works rolmstly and currently has a throughput of about 98% on Verbmobil tree bank input (i.e. it general:es 21,222 English and 26,789 German VITx.) The remaining 2% are due to errors in the SC module, errors in the tree bank, or coordination 1)tel)ictus 1)etween SC and tree 1)ank.</Paragraph>
    <Paragraph position="1"> Evaluation of the module is COnll)lieated 1) 3, the effort involved in mmmally constructing a sizable set of input structures and correspondillg semantic rel)resentations, l?urthermore, t;t1(: VIT formalism has been in constant flux over the last; years with the correct outlmt representations changing almost monthly. It is, however, envisaged to perform an evaluation on('e dust has settled.</Paragraph>
    <Paragraph position="2"> The approach described adds in two respects to the rot)ustnexs of the overall system. First, l;he flat analysis lmrxerx used are very rolmst as conterns low-level inconsistencies such as agreement failures or missing function words (prepositions, determiners, COmlflementizers, etc.). Second, the data analysed in the Verbmobil tree banks m:e exclusively xpoken language. Hence, the tree banks encode analyxes tbr phenomena such as fl'agmentary inlmt , truncated or elliptical selltencex, etc. The described module, gives semantic analyses for all of these constructions. (Usually an abstract predicate is incorl)orated which given a hint to subsequent modules that aim to piece together partial utterances.) Another perspective of this work is that it provides a first step towards a real corlms semantics by converting large sets of data into semantic representations. Due to the abstraction they embody, semantic representations are a valuable too\] for content queries to the. processed corpora.</Paragraph>
    <Paragraph position="3"> More immediately, the semmltic representations 4More inibrmation can be found in Schiehlen (1999).</Paragraph>
    <Paragraph position="4"> generated by the described module have been used as text and training data tbr applications requiring abstract input, such as transfer in machine translation and generation.</Paragraph>
  </Section>
class="xml-element"></Paper>