<?xml version="1.0" standalone="yes"?>
<Paper uid="W01-1404">
  <Title>Approximating Context-Free by Rational Transduction for Example-Based MT</Title>
  <Section position="8" start_page="0" end_page="0" type="concl">
    <SectionTitle>
7 Conclusions
</SectionTitle>
    <Paragraph position="0"> For our application, context-free transduction has a relatively high accuracy, but it also has a high time consumption, and it may be difficult to obtain robustness without further increasing the time costs. These are two major obstacles for use in spoken language systems. We have tried to obtain a rational transduction that approximates a 4It uses a trie to represent productions (similar to ELR parsing (Nederhof, 1994)), postponing generation of output for a production until all nonterminals and all input symbols from the right-hand side have been found.</Paragraph>
    <Paragraph position="1">  context-free transduction, preserving some of its accuracy.</Paragraph>
    <Paragraph position="2"> Our experiments show that the automata we obtain become very large for training corpora of increasing sizes. This poses a problem for determinization. We conjecture that the main source of the excessive growth of the automata lies in noise in the bitexts and their hierarchical alignments. It is a subject for further study whether we can reduce the impact of this noise, e.g. by clustering of source symbols, or by removing some infrequent, idiosyncratic rules from the obtained transduction grammar. Also, other methods of regular approximation of context-free grammars may be considered. null In comparison to a simpler model, viz. bigrams, our approximating transductions do not have a very high accuracy, which is especially worrying since the off-line costs of computation are much higher than in the case of bigrams. The relatively low accuracy may be due to sparseness of data when attaching weights to transitions: the size of the minimal deterministic automaton grows much faster than the size of the training corpus it is constructed from, and the same training corpus is used to train the weights of the transitions of the automaton. Thereby, many transitions do not obtain accurate weights, and unseen input sentences are not translated accurately.</Paragraph>
    <Paragraph position="3"> The problems described here may be avoided by leaving out the determinization of the automaton. This however leads to two new problems: training of the weights requires more sophisticated algorithms, and we may expect an increase in the time needed to transduce input sentences, since now both source and target symbols give  rise to nondeterminism. Whether these problems can be overcome requires further study.</Paragraph>
  </Section>
class="xml-element"></Paper>