<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<item>
  <id>01962764</id>
  <dt>a</dt>
  <an>01962764</an>
  <augroup>
    <au>Allauzen, Cyril</au>
    <au>Mohri, Mehryar</au>
  </augroup>
  <ti>$p$-subsequentiable transducers.</ti>
  <so>Champarnaud, Jean-Marc (ed.) et al., Implementation and application of automata. 7th international conference, CIAA 2002, Tours, France, July 3--5, 2002. Revised papers. Berlin: Springer (ISBN 3-540-40391-4/pbk). Lect. Notes Comput. Sci. 2608, 24-34 (2003).</so>
  <py>2003</py>
  <pu>Berlin: Springer</pu>
  <lagroup>
    <la>EN</la>
  </lagroup>
  <ccgroup>
  </ccgroup>
  <utgroup>
  </utgroup>
  <cigroup>
  </cigroup>
  <ligroup>
    <li>http://link.springer.de/link/service/series/0558/bibs/2608/26080024.htm</li>
  </ligroup>
  <abgroup>
    <ab>Summary: $p$-subsequential transducers are efficient finite-state transducers with $p$ final outputs used in a variety of applications. Not all transducers admit equivalent $p$-subsequential transducers however. We briefly describe an existing generalized determinization algorithm for $p$-subsequential transducers and give the first characterization of $p$-subsequ-entiable transducers, transducers that admit equivalent $p$-subsequential transducers. Our characterization shows the existence of an efficient algorithm for testing $p$-subsequentiability. We have fully implemented the generalized determinization algorithm and the algorithm for testing $p$-subsequentiability. We report experimental results showing that these algorithms are practical in large-vocabulary speech recognition applications. The theoretical formulation of our results is the equivalence of the following three properties for finite-state transducers: determinizability in the sense of the generalized algorithm, $p$-subsequentiability, and the twins property.</ab>
    <rv></rv>
  </abgroup>
</item>