<?xml version="1.0" standalone="yes"?>
<Paper uid="N01-1015">
  <Title>Re-Engineering Letter-to-Sound Rules</Title>
  <Section position="9" start_page="0" end_page="0" type="concl">
    <SectionTitle>
7 Conclusion
</SectionTitle>
    <Paragraph position="0"> We presented a method for building efficient letter-to-sound rules from information extractable from, or with the help of, existing hand-crafted rewrite rules. Using decision trees as the new target representation, significant improvements in time and space efficiency could be achieved at the cost of a reduction in accuracy. Our approach relies on finding an alignment between strings of letters and phonemes. We identified a way to improve alignments and argued that finding a good alignment is crucial for success and should receive more attention. null</Paragraph>
  </Section>
class="xml-element"></Paper>