... Proceedings of ACL-08: HLT, pages 97–105,
Columbus, Ohio, USA, June 2008.
c
2008 Association for Computational Linguistics
Bayesian Learning of Non-compositional Phrases with Synchronous Parsing
Hao ... 14627
gildea@cs.rochester.edu
Abstract
We combine the strengths of Bayesian mod-
eling and synchronous grammar in unsu-
pervised learning of basic translation phrase
pai...
... the performance of that grammar to that
of a heuristically pruned “minimal subset” of it.
The latter’s performance was quite good, achiev-
ing 90.8% F
1
score
1
on section 23 of the WSJ.
This ... training data is a set of
parse trees T that we assume was produced by an
unknown TSG g with probability Pr(T |g). Using
Bayes’ rule, we can compute the probability of a
particular hypo...
... Linguistics
Automatic learning of textual entailments with cross-pair similarities
Fabio Massimo Zanzotto
DISCo
University of Milano-Bicocca
Milan, Italy
zanzotto@disco.unimib.it
Alessandro Moschitti
Department of ... rules that describe a non trivial set
of entailment cases. The experiments with
the data sets of the RTE 2005 challenge
show an improvement of 4.4% over the
st...
...
built with and without PW. In Table 2, the PW-
prefix refers to the DLMs with PW = 0.5, and the
DLMs without PW- prefix refers to DLMs with PW
= 0. For both DLM_1 and DLM_2, models with the ... 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 in Table 1, i...
... Proceedings of the 47th Annual Meeting of the ACL and the 4th IJCNLP of the AFNLP, pages 602–610,
Suntec, Singapore, 2-7 August 2009.
c
2009 ACL and AFNLP
Unsupervised Learning of Narrative ... role of protagonist. Chain learning and clus-
tering is based only on the frequency with which
two verbs share arguments, ignoring any features
of the arguments themselves.
Take this...
... dataset of 597 examples.
The data is labeled with two different sets of
semantic relations: one set of 30 relations with
fairly specific meanings, and another set of 5 rela-
tions with more ... this paper was twofold. Firstly,
we wanted to compare the performance of different
machine learning algorithms on the task of map-
ping from a vector of web frequencies of...
... increase in
recall of nearly 3% absolute with a slight drop in
precision. These results are very promising and
further show the robustness of discriminative on-
line learning with approximate parsing ... algorithms within
an online learning framework, which has been
shown to be robust with respect approximate in-
ference, and describe experiments displaying that
these new models...
... France
emmanuel.dupoux@gmail.com
Abstract
Accurate unsupervised learning of phonemes
of a language directly from speech is demon-
strated via an algorithm for joint unsupervised
learning of the topology and parameters of
a hidden Markov ... The
algorithm, originally proposed for unsuper-
vised learning of allophonic variations within
a given phoneme set, has been adapted to
le...
... two examples use the joint probability
of the prefix and suffix, with a smoothing back-off
(the product of the individual probabilities). Scor-
ing models of this form proved to be poor perform-
ers ... additional
challenge of infixes, we did not tackle them because
they often substantially change the meaning. Irregu-
lar morphology is also beyond the scope of this pa-
per. As a side...
... proofs of theorems 1 and 2. Due
to space limitations we cannot give full proofs; in-
stead we provide proofs of some key lemmas. A
long version of this paper will give the full proofs.
9.1 Proof ... the conven-
tional form of the inside-outside algorithm.
The proof is by induction, and is similar to the
proof of lemma 2; for reasons of space it is omitted.
9.2 Proof of the Ident...