<?xml version="1.0" standalone="yes"?>
<Paper uid="P98-2223">
  <Title>A Pattern-based Machine Translation System Extended by Example-based Processing</Title>
  <Section position="6" start_page="1371" end_page="1371" type="evalu">
    <SectionTitle>
5 Discussion
</SectionTitle>
    <Paragraph position="0"> Unlike most of existing MT approaches that consist of three major components\[I, 2\] - analysis, transfer, and generation - the pattern-based MT is based on a synchronous model\[5, 3\] of translation. That is, the analysis of a source sentence is directly connected to the generation of a target sentence through the translation knowledge (i.e., patterns). This simple architecture makes it much easier to customize a system for improving translation quality than the conventional MT, since the management of the ambiguities in 3-component architecture has to tackle the exponential combination of overall ambiguities. In this simple model, we can concentrate on a single module (a parser with synchronous derivation), and manage most of translation knowledge in a uniform way as translation patterns.</Paragraph>
    <Paragraph position="1"> Although it is easier to add translation patterns in our system than previous systems, it is difficult for non-experts to specify detailed matching conditions (or features). Therefore, we made a pattern compiler which interprets a simple pattern which a non-expert writes and converts it into the full-scale patterns including necessary matching conditions,  Shown in the above example, it is very easy for non-experts to write these simple patterns. Thus, this pattern compiler enable non-experts to customize a system. In conventional MT systems, an expert is usually needed for each component (analysis, transfer, and generation).</Paragraph>
    <Paragraph position="2"> These advantages can reduce the cost of development and customization of a MT system, and can largely contribute to rapidly improve the translation quality in a short time.</Paragraph>
    <Paragraph position="3"> Further, we have shown the way to integrate example-based processing and pattern-based MT.</Paragraph>
    <Paragraph position="4"> In addition to reduce the total number of translation patterns, this combination enables us to make a more robust and human-like MT system thanks to the easy addition of translation pattern.</Paragraph>
  </Section>
class="xml-element"></Paper>