corpus </term> . We conducted experiments with an <term> EBMT system </term> . The two <term>
a <term> node </term> in a <term> graph </term> with <term> in-degree </term> greater than one and
<term> planning-based architecture </term> with a variety of <term> language processing modules
dialogue </term> . Our <term> system </term> deals with <term> pronouns </term> with <term> NP - and
system </term> deals with <term> pronouns </term> with <term> NP - and non-NP-antecedents </term>
</term> results in 87.5 % <term> agreement </term> with a state of the art , proprietary <term> Arabic
re-estimate the <term> model parameters </term> with the expanded <term> vocabulary </term> and <term>
conducted <term> psychological experiments </term> with 42 subjects to collect <term> referring expressions
classification ( maximum entropy ) </term> with <term> linguistic information </term> . Instead
According to our assumption , most of the words with similar <term> context features </term> in
parametric affinity model </term> . In comparison with previous <term> models </term> , which either
is scalable . We apply it in combination with a <term> terabyte corpus </term> to answer <term>
sentences </term> in student <term> writing </term> with <term> essay-based discourse elements </term>
Intra-sentential quality </term> is evaluated with <term> rule-based heuristics </term> . Results
of <term> words </term> that tend to co-occur with it . <term> Topic signatures </term> can be
</term> on standard <term> WSD datasets </term> , with promising results . This paper presents
of the <term> algorithm </term> is contrasted with <term> human annotation performance </term>
phrases </term> , i.e. <term> phrases </term> with gaps . A <term> method </term> for producing
translation accuracy </term> , as measured with the <term> NIST evaluation metric </term> .
<term> computational problems </term> associated with <term> probabilistic translation models </term>
hide detail