... Proceedings of the 43rd Annual Meeting of the ACL, pages 26–33,
Ann Arbor, June 2005.
c
2005 Association for Computational Linguistics
Supersense Tagging of Unknown Nouns using Semantic Similarity
James ... which can be alleviated by automati-
cally classifying the unknown words. Su-
persense tagging assigns unknown nouns one
of 26 broad semantic categories used b...
... performance of each single
ad (its persuasiveness, from our point of view):
• CTR, Click Through Rate: measures the num-
ber of clicks divided by the number of impres-
sions (i.e. the number of times ... experimentally exploring
other aspects of the wide field of pragmatics.
The paper is structured as follows: Section 2 dis-
cusses the main advantages of ecological approaches...
... indicates an
improvement of 22-38% in average pre-
cision over unstemmed text, and 96% of
the performance of the proprietary stem-
mer above.
1 Introduction
Stemming is the process of normalizing word ... phase was adapted
to use the sum of the weights of a word in a sentence
instead of the count.
2.3 Candidate Stem Scoring
As previously mentioned, each word has a list of
subs...
... 0
Table 2. Distribution of phenotypes of erythrocyte histone H1.a in progeny from various matings of Rhode Island Red R11 chickens.
Type of mating Number of
Frequency of phenotypes in progeny
(observed ... arrays of nucleosomes are linked
together by DNA to form chromatin fibers with a
diameter of 10 nm. In bulk chromatin, one molecule
of histone H1 binds to a linker DNA seal...
... amplitude of the
fast phase of the l atter represen ted up to 40% of the total
amplitude, with a rate constant (k
obs
) independent of Pc
concentration (data not shown). From the k
obs
values of this
fast ... directly
with the pH gradient in the chromatofocusing step, avo iding
the l ast salt gradient. About 3 mg of pure Pc with an
absorba nce ratio A
275
/A
590
of 1.27 were obt...
...
is then rescored using a more sophisticated lan-
guage model. We use the N-best list of N=100,
whose “oracle” CER (i.e., the CER of the hy-
potheses with the minimum number of errors) is
presented ... number of pa-
rameters of the combined model manageable. To
overcome the second obstacle, we used an unsu-
pervised learning method that discovers the de-
pendency structure of...
... because of unknown words,
style and frequently used grammatical structures).
As the fiction subcorpus consists of three books
and the law subcorpus consists of two rules, we
took half of one of the ... Yoram Singer. 2003. Feature-rich part-
of- speech tagging with a cyclic dependency net-
work. In Proceedings of the 2003 Conference
of the North American Chapter of the Asso...
... conducted in the pres-
ence of saturating concentrations of dUMP, to ensure
examination of the dUMP-activated form of the
enzyme [12,13]. The results revealed that, in the
absence of CH
2
H
4
folate, ... concentrations of
dUMP (0–200 lm) . Reactions were initiated by addition of
NADPH stock solution. The final volume of the reaction
mixture was 300 lL. The reduction of FAD an...
... performed in the presence of the DMC1 (6 lM) or DMC1-I37N (6 lM)
protein in the same manner as described in (B), except for the presence of the indicated amounts of CaCl
2
.
Activity of the human DMC1-I37N ... polymorphism may have the potential to
be a source of human infertility. Extensive SNP analyses
of the DMC1-I37N polymorphism in the human popu-
lation and in sets of infe...
... document frequency (IDF) of
term
w
(up to bigram), which was calculated as the
logarithm of the ratio of the number of articles that
w
appears over the total number of all articles in the
document ... is
that of Problem 1.
3.3 Automatic summarization evaluation
Automatic evaluation of summary quality is impor-
tant for the research of document summarization as
it avoids the la...