|
performance gains from the
<term>
data
</term>
by
|
using
|
<term>
class-dependent interpolation
</term>
|
#3073
In this paper, we show how training data can be supplemented with text from the web filtered to match the style and/or topic of the target recognition task, but also that it is possible to get bigger performance gains from the data by using class-dependent interpolation of N-grams. |
|
Path-based inference rules
</term>
may be written
|
using
|
a
<term>
binary relational calculus notation
|
#12099
Path-based inference rules may be written using a binary relational calculus notation. |
|
</term>
based on the results . The evaluation
|
using
|
another 23 subjects showed that the proposed
|
#5713
The evaluation using another 23 subjects showed that the proposed method could effectively generate proper referring expressions. |
|
sense disambiguation performance
</term>
,
|
using
|
standard
<term>
WSD evaluation methodology
|
#7814
We present the first known empirical test of an increasingly common speculative claim, by evaluating a representative Chinese-to-English SMT model directly on word sense disambiguation performance, using standard WSD evaluation methodology and datasets from the Senseval-3 Chinese lexical sample task. |
|
a
<term>
token classification task
</term>
,
|
using
|
various
<term>
tagging strategies
</term>
to
|
#10813
There are several approaches that model information extraction as a token classification task, using various tagging strategies to combine multiple tokens. |
|
for
<term>
Japanese sentence analyses
</term>
|
using
|
an
<term>
argumentation system
</term>
by Konolige
|
#16579
This paper proposes that sentence analysis should be treated as defeasible reasoning, and presents such a treatment for Japanese sentence analysesusing an argumentation system by Konolige, which is a formalization of defeasible reasoning, that includes arguments and defeat rules that capture defeasibility. |
|
<term>
phrases
</term>
while simultaneously
|
using
|
less
<term>
memory
</term>
than is required
|
#9146
In this paper we describe a novel data structure for phrase-based statistical machine translation which allows for the retrieval of arbitrarily long phrases while simultaneously using less memory than is required by current decoder implementations. |
|
describe the methods and hardware that we are
|
using
|
to produce a real-time demonstration of
|
#16874
We describe the methods and hardware that we are using to produce a real-time demonstration of an integrated Spoken Language System. |
|
In this paper , we describe the research
|
using
|
<term>
machine learning techniques
</term>
|
#11208
In this paper, we describe the research using machine learning techniques to build a comma checker to be integrated in a grammar checker for Basque. |
|
systems still treat
<term>
coordination
</term>
|
using
|
adapted
<term>
parsing strategies
</term>
,
|
#21150
Despite the large amount of theoretical work done on non-constituent coordination during the last two decades, many computational systems still treat coordinationusing adapted parsing strategies, in a similar fashion to the SYSCONJ system developed for ATNs. |
|
Sentence ambiguities
</term>
can be resolved by
|
using
|
domain targeted preference knowledge without
|
#16325
Sentence ambiguities can be resolved by using domain targeted preference knowledge without using complicated large knowledgebases. |
|
ambiguities
</term>
by indirectly and implicitly
|
using
|
<term>
maximum likelihood method
</term>
,
|
#17849
Owing to the problem of insufficient training data and approximation error introduced by the language model, traditional statistical approaches, which resolve ambiguities by indirectly and implicitly using maximum likelihood method, fail to achieve high performance in real applications. |
|
outputs
</term>
of our
<term>
MT system
</term>
|
using
|
the
<term>
NIST and Bleu automatic MT evaluation
|
#9517
We evaluate the outputs of our MT systemusing the NIST and Bleu automatic MT evaluation software. |
|
<term>
word string
</term>
has been obtained by
|
using
|
a different
<term>
LM
</term>
. Actually ,
|
#1110
The oracle knows the reference word string and selects the word string with the best performance (typically, word or semantic error rate) from a list of word strings, where each word string has been obtained by using a different LM. |
|
models
</term>
. The models were constructed
|
using
|
a 5K
<term>
vocabulary
</term>
and trained
|
#21246
The models were constructed using a 5K vocabulary and trained using a 76 million word Wall Street Journal text corpus. |
|
corpora
</term>
is presented which involves
|
using
|
a
<term>
statistical POS tagger
</term>
in
|
#19958
A novel method for adding linguistic annotation to corpora is presented which involves using a statistical POS tagger in conjunction with unsupervised structure finding methods to derive notions of noun group, verb group, and so on which is inherently extensible to more sophisticated annotation, and does not require a pre-tagged corpus to fit. |
|
automatically from
<term>
raw text
</term>
. Experiments
|
using
|
the
<term>
SemCor
</term>
and
<term>
Senseval-3
|
#11025
Experiments using the SemCor and Senseval-3 data sets demonstrate that our ensembles yield significantly better results when compared with state-of-the-art. |
|
efficient
<term>
decoder
</term>
and show that
|
using
|
these
<term>
tree-based models
</term>
in combination
|
#9281
We describe an efficient decoder and show that using these tree-based models in combination with conventional SMT models provides a promising approach that incorporates the power of phrasal SMT with the linguistic generality available in a parser. |
|
segments of actual tape-recorded descriptions ,
|
using
|
<term>
organizational and discourse strategies
|
#15485
The model is embodied in a program, APT, that can reproduce segments of actual tape-recorded descriptions, using organizational and discourse strategies derived through analysis of our corpus. |
|
parser
</term>
skips that
<term>
portion
</term>
|
using
|
a fake
<term>
non-terminal symbol
</term>
.
|
#18182
When at very noisy portion is detected, the parser skips that portionusing a fake non-terminal symbol. |