... each ofthe 21 nouns. The sense
with the highest estimated sense prior is taken as the
predominant senseofthe noun.
For the set of 12 nouns where the predominant
54
Proceedings ofthe 45th Annual ... predominant senseofthe noun interest
in the BC part ofthe DSO corpus has the meaning
“a senseof concern with and curiosity about some-
one or something”. In the WSJ part ofthe DSO cor-
pus, the ... one
of the reasons forthe drop in accuracy is the dif-
ference in sense priors (i.e., the proportions of the
different senses of a word) between BC and WSJ.
When the authors assumed they knew the...
... L., and Palmer, M. 2006.
An EmpiricalStudyoftheBehaviorofActive
Learning forWordSense Disambiguation, Proc. of
the main conference on Human Language Tech-
nology Conference ofthe ... num-
ber of its senses, the number of its data instances,
the number of feature, and the percentage of
positive sense instances for each data set.
Assigning the correct labels of data instances ... forword
sense disambiguation (WSD) in the do-
main of web queries, where a complete set
of ambiguous word senses are unknown.
In this paper, we present a combination of
active learning and...
... accuracy) and cost of annotating a sentence
depend not only on properties ofthe sentence but
also on the order in which the items are annotated.
Therefore, when evaluating the performance of an
AL ... less
human effort.
Annotation cost is project dependent. For in-
stance, annotators may be paid forthe number of an-
notations they produce or by the hour. In the context
of parse tree annotation, ... employed to reduce the costs of corpus annotation
(Engelson and Dagan, 1996; Ringger et al., 2007;
Tomanek et al., 2007). With the assistance of AL,
the role ofthe human oracle is either to label...
... which
lists seven senses forthe noun channel. Two
senses are lumped together if they are translated in
the same way in Chinese. For example, sense 1 and
7 of channel are both translated as “频道” ... Evaluating WordSense
Disambiguation Systems (SENSEVAL-2), pages 1-5.
Gerard Escudero, Lluis Marquez, and German Rigau.
2000. Anempiricalstudyofthe domain dependence
of supervised wordsensedisambiguation ... classifier to determine
the most probable senseof w.
of the senses of some nouns. For instance, no oc-
currences ofsense 5 ofthe noun circuit (racing
circuit, a racetrack for automobile races)...
... each sense
of one ofthe words. Pick one ofthe words,
say W2, and using WordNet, form a similarity
list for each senseof that word. For this, use
the words from the synset of each sense and ... one of these queries,
we get the number of hits for each sense i of W2
and this provides a ranking ofthe m senses of
W2 as they relate with 1411.
Example The types of query that can be formed ... summation ofthe conceptual densities be-
tween thesense i oftheword X and all the
senses ofthe words Y. The results are shown
in the tables below where the conceptual den-
sity calculated for...
... own performance on
an alternative layout. The practice diagrams and the
randomisation ofthe order of presentation of the
experimental diagrams for each subject helped counter
the learning effect ... related to the nature ofthe task and the form of the
experimental materials. Students said that they found the
diagrams easier to understand if, when reading from top
to bottom, the order ofthe classes ... 1994): they take as input a relational graph structure
of objects and the relationships between them, and
produce a visual representation ofthe information in
diagrammatic form. The designers of these...
... houses in the Hot List
3
is randomly assigned to one ofthe three
conditions.
Then, the subject interacts with the evaluation
framework and at the end ofthe interaction
measures ofthe argument ... Figure 1 for a simple value tree in
the real estate domain). The arcs ofthe tree are
weighted to represent the importance ofthe
value ofan objective in contributing to the value
of its parent ... explicitly questioning the user at the end of
the interaction about the rationale for her
decision (Olso and Zanna 1991). This can
provide valuable information on what aspects of
the argument were...
... A set ofthe attributes of
,
and
is used to predict the label ofthe . The
set consists of twenty attributes: ten forthe char-
acter type (
, , , ,
, , , , , ), and an-
other ten forthe character ... 500 and 1000.
However, they have concluded that a larger pool is better than
a smaller one because the final accuracy ofthe former is higher
than that ofthe latter.
6
The variance of a set of ... AnEmpiricalStudyofActiveLearning with Support Vector Machines for
Japanese Word Segmentation
Manabu Sassano
Fujitsu Laboratories Ltd.
4-1-1, Kamikodanaka, Nakahara-ku,
Kawasaki...
... Pro-
cessing Conference and the 1st Conference of the
North American Chapter ofthe Association for
Computational Linguistics, Seattle, WA, April.
An EmpiricalStudyof Information Synthesis Tasks
Enrique ... selected the Spanish CLEF 2001-2003
news collection testbed (Peters et al., 2002), be-
cause Spanish is the native language ofthe subjects
recruited forthe manual generation of reports. Out
of the ... substantially bet-
ter than ROUGE for a relevant class of topics.
Section 3 describes these metrics and the experi-
mental design to compare them; in Section 4, we an-
alyze the outcome of the...
... we compare
the performance ofthe state -of- the- art ma-
chine learning models. Then we propose
two approaches in order to improve the
performance of Chinese chunking. 1) We
propose an approach ... bi-grams of words in
an n window.
• POS: uni-gram and bi-grams of POS in an n
window.
• WORD+ POS: Both the features of WORD
and POS.
where n is a predefined number to denote window
size.
For instance, ... WORD+ POS, ”P” refers to POS. We can
see from the figure that WORD+ POS yielded bet-
ter performance than POS in the most cases. How-
ever, when the size of training data was small,
the performance...
... This reason can be considered causes ofthe financial crisis, and the
subsequent events of corporate collapses and accounting fraud.
An organization lacks of transparency in their financial report ... In the United State, after the bankruptcy of Enron and Worldcom, the
- 40 -
reason can bring a not good sufficient for them to present all the view of accountants
in Hanoi. In additional, the ... mainly on the accounting information
and financial statements. Therefore it is the responsibility ofthe accounting
professionals to stick to the code of ethics, i.e. this is the role of ethics...