Titles
stringlengths 6
220
| Abstracts
stringlengths 37
3.26k
| Years
int64 1.99k
2.02k
| Categories
stringclasses 1
value |
---|---|---|---|
Challenges for Distributional Compositional Semantics | This paper summarises the current state-of-the art in the study of
compositionality in distributional semantics, and major challenges for this
area. We single out generalised quantifiers and intensional semantics as areas
on which to focus attention for the development of the theory. Once suitable
theories have been developed, algorithms will be needed to apply the theory to
tasks. Evaluation is a major problem; we single out application to recognising
textual entailment and machine translation for this purpose.
| 2,012 | Computation and Language |
Distinct word length frequencies: distributions and symbol entropies | The distribution of frequency counts of distinct words by length in a
language's vocabulary will be analyzed using two methods. The first, will look
at the empirical distributions of several languages and derive a distribution
that reasonably explains the number of distinct words as a function of length.
We will be able to derive the frequency count, mean word length, and variance
of word length based on the marginal probability of letters and spaces. The
second, based on information theory, will demonstrate that the conditional
entropies can also be used to estimate the frequency of distinct words of a
given length in a language. In addition, it will be shown how these techniques
can also be applied to estimate higher order entropies using vocabulary word
length.
| 2,012 | Computation and Language |
Clustering based approach extracting collocations | The following study presents a collocation extraction approach based on
clustering technique. This study uses a combination of several classical
measures which cover all aspects of a given corpus then it suggests separating
bigrams found in the corpus in several disjoint groups according to the
probability of presence of collocations. This will allow excluding groups where
the presence of collocations is very unlikely and thus reducing in a meaningful
way the search space.
| 2,012 | Computation and Language |
Automatic Segmentation of Manipuri (Meiteilon) Word into Syllabic Units | The work of automatic segmentation of a Manipuri language (or Meiteilon) word
into syllabic units is demonstrated in this paper. This language is a scheduled
Indian language of Tibeto-Burman origin, which is also a very highly
agglutinative language. This language usages two script: a Bengali script and
Meitei Mayek (Script). The present work is based on the second script. An
algorithm is designed so as to identify mainly the syllables of Manipuri origin
word. The result of the algorithm shows a Recall of 74.77, Precision of 91.21
and F-Score of 82.18 which is a reasonable score with the first attempt of such
kind for this language.
| 2,012 | Computation and Language |
Frame Interpretation and Validation in a Open Domain Dialogue System | Our goal in this paper is to establish a means for a dialogue platform to be
able to cope with open domains considering the possible interaction between the
embodied agent and humans. To this end we present an algorithm capable of
processing natural language utterances and validate them against knowledge
structures of an intelligent agent's mind. Our algorithm leverages dialogue
techniques in order to solve ambiguities and acquire knowledge about unknown
entities.
| 2,012 | Computation and Language |
Appropriate Nouns with Obligatory Modifiers | The notion of appropriate sequence as introduced by Z. Harris provides a
powerful syntactic way of analysing the detailed meaning of various sentences,
including ambiguous ones. In an adjectival sentence like 'The leather was
yellow', the introduction of an appropriate noun, here 'colour', specifies
which quality the adjective describes. In some other adjectival sentences with
an appropriate noun, that noun plays the same part as 'colour' and seems to be
relevant to the description of the adjective. These appropriate nouns can
usually be used in elementary sentences like 'The leather had some colour', but
in many cases they have a more or less obligatory modifier. For example, you
can hardly mention that an object has a colour without qualifying that colour
at all. About 300 French nouns are appropriate in at least one adjectival
sentence and have an obligatory modifier. They enter in a number of sentence
structures related by several syntactic transformations. The appropriateness of
the noun and the fact that the modifier is obligatory are reflected in these
transformations. The description of these syntactic phenomena provides a basis
for a classification of these nouns. It also concerns the lexical properties of
thousands of predicative adjectives, and in particular the relations between
the sentence without the noun : 'The leather was yellow' and the adjectival
sentence with the noun : 'The colour of the leather was yellow'.
| 1,995 | Computation and Language |
A prototype for projecting HPSG syntactic lexica towards LMF | The comparative evaluation of Arabic HPSG grammar lexica requires a deep
study of their linguistic coverage. The complexity of this task results mainly
from the heterogeneity of the descriptive components within those lexica
(underlying linguistic resources and different data categories, for example).
It is therefore essential to define more homogeneous representations, which in
turn will enable us to compare them and eventually merge them. In this context,
we present a method for comparing HPSG lexica based on a rule system. This
method is implemented within a prototype for the projection from Arabic HPSG to
a normalised pivot language compliant with LMF (ISO 24613 - Lexical Markup
Framework) and serialised using a TEI (Text Encoding Initiative) based
representation. The design of this system is based on an initial study of the
HPSG formalism looking at its adequacy for the representation of Arabic, and
from this, we identify the appropriate feature structures corresponding to each
Arabic lexical category and their possible LMF counterparts.
| 2,012 | Computation and Language |
FST Based Morphological Analyzer for Hindi Language | Hindi being a highly inflectional language, FST (Finite State Transducer)
based approach is most efficient for developing a morphological analyzer for
this language. The work presented in this paper uses the SFST (Stuttgart Finite
State Transducer) tool for generating the FST. A lexicon of root words is
created. Rules are then added for generating inflectional and derivational
words from these root words. The Morph Analyzer developed was used in a Part Of
Speech (POS) Tagger based on Stanford POS Tagger. The system was first trained
using a manually tagged corpus and MAXENT (Maximum Entropy) approach of
Stanford POS tagger was then used for tagging input sentences. The
morphological analyzer gives approximately 97% correct results. POS tagger
gives an accuracy of approximately 87% for the sentences that have the words
known to the trained model file, and 80% accuracy for the sentences that have
the words unknown to the trained model file.
| 2,012 | Computation and Language |
Adaptation of pedagogical resources description standard (LOM) with the
specificity of Arabic language | In this article we focus firstly on the principle of pedagogical indexing and
characteristics of Arabic language and secondly on the possibility of adapting
the standard for describing learning resources used (the LOM and its
Application Profiles) with learning conditions such as the educational levels
of students and their levels of understanding,... the educational context with
taking into account the representative elements of text, text length, ... in
particular, we put in relief the specificity of the Arabic language which is a
complex language, characterized by its flexion, its voyellation and
agglutination.
| 2,012 | Computation and Language |
A Method for Selecting Noun Sense using Co-occurrence Relation in
English-Korean Translation | The sense analysis is still critical problem in machine translation system,
especially such as English-Korean translation which the syntactical different
between source and target languages is very great. We suggest a method for
selecting the noun sense using contextual feature in English-Korean
Translation.
| 2,012 | Computation and Language |
More than Word Frequencies: Authorship Attribution via Natural Frequency
Zoned Word Distribution Analysis | With such increasing popularity and availability of digital text data,
authorships of digital texts can not be taken for granted due to the ease of
copying and parsing. This paper presents a new text style analysis called
natural frequency zoned word distribution analysis (NFZ-WDA), and then a basic
authorship attribution scheme and an open authorship attribution scheme for
digital texts based on the analysis. NFZ-WDA is based on the observation that
all authors leave distinct intrinsic word usage traces on texts written by them
and these intrinsic styles can be identified and employed to analyze the
authorship. The intrinsic word usage styles can be estimated through the
analysis of word distribution within a text, which is more than normal word
frequency analysis and can be expressed as: which groups of words are used in
the text; how frequently does each group of words occur; how are the
occurrences of each group of words distributed in the text. Next, the basic
authorship attribution scheme and the open authorship attribution scheme
provide solutions for both closed and open authorship attribution problems.
Through analysis and extensive experimental studies, this paper demonstrates
the efficiency of the proposed method for authorship attribution.
| 2,012 | Computation and Language |
Recent Technological Advances in Natural Language Processing and
Artificial Intelligence | A recent advance in computer technology has permitted scientists to implement
and test algorithms that were known from quite some time (or not) but which
were computationally expensive. Two such projects are IBM's Jeopardy as a part
of its DeepQA project [1] and Wolfram's Wolframalpha[2]. Both these methods
implement natural language processing (another goal of AI scientists) and try
to answer questions as asked by the user. Though the goal of the two projects
is similar, both of them have a different procedure at it's core. In the
following sections, the mechanism and history of IBM's Jeopardy and Wolfram
alpha has been explained followed by the implications of these projects in
realizing Ray Kurzweil's [3] dream of passing the Turing test by 2029. A recipe
of taking the above projects to a new level is also explained.
| 2,012 | Computation and Language |
Introduction of the weight edition errors in the Levenshtein distance | In this paper, we present a new approach dedicated to correcting the spelling
errors of the Arabic language. This approach corrects typographical errors like
inserting, deleting, and permutation. Our method is inspired from the
Levenshtein algorithm, and allows a finer and better scheduling than
Levenshtein. The results obtained are very satisfactory and encouraging, which
shows the interest of our new approach.
| 2,012 | Computation and Language |
Average word length dynamics as indicator of cultural changes in society | Dynamics of average length of words in Russian and English is analysed in the
article. Words belonging to the diachronic text corpus Google Books Ngram and
dated back to the last two centuries are studied. It was found out that average
word length slightly increased in the 19th century, and then it was growing
rapidly most of the 20th century and started decreasing over the period from
the end of the 20th - to the beginning of the 21th century. Words which
contributed mostly to increase or decrease of word average length were
identified. At that, content words and functional words are analysed
separately. Long content words contribute mostly to word average length of
word. As it was shown, these words reflect the main tendencies of social
development and thus, are used frequently. Change of frequency of personal
pronouns also contributes significantly to change of average word length. The
other parameters connected with average length of word were also analysed.
| 2,015 | Computation and Language |
Authorship Identification in Bengali Literature: a Comparative Analysis | Stylometry is the study of the unique linguistic styles and writing behaviors
of individuals. It belongs to the core task of text categorization like
authorship identification, plagiarism detection etc. Though reasonable number
of studies have been conducted in English language, no major work has been done
so far in Bengali. In this work, We will present a demonstration of authorship
identification of the documents written in Bengali. We adopt a set of
fine-grained stylistic features for the analysis of the text and use them to
develop two different models: statistical similarity model consisting of three
measures and their combination, and machine learning model with Decision Tree,
Neural Network and SVM. Experimental results show that SVM outperforms other
state-of-the-art methods after 10-fold cross validations. We also validate the
relative importance of each stylistic feature to show that some of them remain
consistently significant in every model used in this experiment.
| 2,012 | Computation and Language |
Robopinion: Opinion Mining Framework Inspired by Autonomous Robot
Navigation | Data association methods are used by autonomous robots to find matches
between the current landmarks and the new set of observed features. We seek a
framework for opinion mining to benefit from advancements in autonomous robot
navigation in both research and development
| 2,012 | Computation and Language |
Input Scheme for Hindi Using Phonetic Mapping | Written Communication on Computers requires knowledge of writing text for the
desired language using Computer. Mostly people do not use any other language
besides English. This creates a barrier. To resolve this issue we have
developed a scheme to input text in Hindi using phonetic mapping scheme. Using
this scheme we generate intermediate code strings and match them with
pronunciations of input text. Our system show significant success over other
input systems available.
| 2,012 | Computation and Language |
Evaluation of Computational Grammar Formalisms for Indian Languages | Natural Language Parsing has been the most prominent research area since the
genesis of Natural Language Processing. Probabilistic Parsers are being
developed to make the process of parser development much easier, accurate and
fast. In Indian context, identification of which Computational Grammar
Formalism is to be used is still a question which needs to be answered. In this
paper we focus on this problem and try to analyze different formalisms for
Indian languages.
| 2,012 | Computation and Language |
Identification of Fertile Translations in Medical Comparable Corpora: a
Morpho-Compositional Approach | This paper defines a method for lexicon in the biomedical domain from
comparable corpora. The method is based on compositional translation and
exploits morpheme-level translation equivalences. It can generate translations
for a large variety of morphologically constructed words and can also generate
'fertile' translations. We show that fertile translations increase the overall
quality of the extracted lexicon for English to French translation.
| 2,012 | Computation and Language |
Stemmer for Serbian language | In linguistic morphology and information retrieval, stemming is the process
for reducing inflected (or sometimes derived) words to their stem, base or root
form; generally a written word form. In this work is presented suffix stripping
stemmer for Serbian language, one of the highly inflectional languages.
| 2,012 | Computation and Language |
Natural Language Processing - A Survey | The utility and power of Natural Language Processing (NLP) seems destined to
change our technological society in profound and fundamental ways. However
there are, to date, few accessible descriptions of the science of NLP that have
been written for a popular audience, or even for an audience of intelligent,
but uninitiated scientists. This paper aims to provide just such an overview.
In short, the objective of this article is to describe the purpose, procedures
and practical applications of NLP in a clear, balanced, and readable way. We
will examine the most recent literature describing the methods and processes of
NLP, analyze some of the challenges that researchers are faced with, and
briefly survey some of the current and future applications of this science to
IT research in general.
| 2,012 | Computation and Language |
A Linguistic Model for Terminology Extraction based Conditional Random
Fields | In this paper, we show the possibility of using a linear Conditional Random
Fields (CRF) for terminology extraction from a specialized text corpus.
| 2,014 | Computation and Language |
Detecting multiword phrases in mathematical text corpora | We present an approach for detecting multiword phrases in mathematical text
corpora. The method used is based on characteristic features of mathematical
terminology. It makes use of a software tool named Lingo which allows to
identify words by means of previously defined dictionaries for specific word
classes as adjectives, personal names or nouns. The detection of multiword
groups is done algorithmically. Possible advantages of the method for indexing
and information retrieval and conclusions for applying dictionary-based methods
of automatic indexing instead of stemming procedures are discussed.
| 2,012 | Computation and Language |
Quick Summary | Quick Summary is an innovate implementation of an automatic document
summarizer that inputs a document in the English language and evaluates each
sentence. The scanner or evaluator determines criteria based on its grammatical
structure and place in the paragraph. The program then asks the user to specify
the number of sentences the person wishes to highlight. For example should the
user ask to have three of the most important sentences, it would highlight the
first and most important sentence in green. Commonly this is the sentence
containing the conclusion. Then Quick Summary finds the second most important
sentence usually called a satellite and highlights it in yellow. This is
usually the topic sentence. Then the program finds the third most important
sentence and highlights it in red. The implementations of this technology are
useful in a society of information overload when a person typically receives 42
emails a day (Microsoft). The paper also is a candid look at difficulty that
machine learning has in textural translating. However, it speaks on how to
overcome the obstacles that historically prevented progress. This paper
proposes mathematical meta-data criteria that justify the place of importance
of a sentence. Just as tools for the study of relational symmetry in
bio-informatics, this tool seeks to classify words with greater clarity.
"Survey Finds Workers Average Only Three Productive Days per Week." Microsoft
News Center. Microsoft. Web. 31 Mar. 2012.
| 2,012 | Computation and Language |
Inference of Fine-grained Attributes of Bengali Corpus for Stylometry
Detection | Stylometry, the science of inferring characteristics of the author from the
characteristics of documents written by that author, is a problem with a long
history and belongs to the core task of Text categorization that involves
authorship identification, plagiarism detection, forensic investigation,
computer security, copyright and estate disputes etc. In this work, we present
a strategy for stylometry detection of documents written in Bengali. We adopt a
set of fine-grained attribute features with a set of lexical markers for the
analysis of the text and use three semi-supervised measures for making
decisions. Finally, a majority voting approach has been taken for final
classification. The system is fully automatic and language-independent.
Evaluation results of our attempt for Bengali author's stylometry detection
show reasonably promising accuracy in comparison to the baseline model.
| 2,011 | Computation and Language |
Opinion Mining for Relating Subjective Expressions and Annual Earnings
in US Financial Statements | Financial statements contain quantitative information and manager's
subjective evaluation of firm's financial status. Using information released in
U.S. 10-K filings. Both qualitative and quantitative appraisals are crucial for
quality financial decisions. To extract such opinioned statements from the
reports, we built tagging models based on the conditional random field (CRF)
techniques, considering a variety of combinations of linguistic factors
including morphology, orthography, predicate-argument structure, syntax, and
simple semantics. Our results show that the CRF models are reasonably effective
to find opinion holders in experiments when we adopted the popular MPQA corpus
for training and testing. The contribution of our paper is to identify opinion
patterns in multiword expressions (MWEs) forms rather than in single word
forms.
We find that the managers of corporations attempt to use more optimistic
words to obfuscate negative financial performance and to accentuate the
positive financial performance. Our results also show that decreasing earnings
were often accompanied by ambiguous and mild statements in the reporting year
and that increasing earnings were stated in assertive and positive way.
| 2,012 | Computation and Language |
Learning Attitudes and Attributes from Multi-Aspect Reviews | The majority of online reviews consist of plain-text feedback together with a
single numeric score. However, there are multiple dimensions to products and
opinions, and understanding the `aspects' that contribute to users' ratings may
help us to better understand their individual preferences. For example, a
user's impression of an audiobook presumably depends on aspects such as the
story and the narrator, and knowing their opinions on these aspects may help us
to recommend better products. In this paper, we build models for rating systems
in which such dimensions are explicit, in the sense that users leave separate
ratings for each aspect of a product. By introducing new corpora consisting of
five million reviews, rated with between three and six aspects, we evaluate our
models on three prediction tasks: First, we use our model to uncover which
parts of a review discuss which of the rated aspects. Second, we use our model
to summarize reviews, which for us means finding the sentences that best
explain a user's rating. Finally, since aspect ratings are optional in many of
the datasets we consider, we use our model to recover those ratings that are
missing from a user's evaluation. Our model matches state-of-the-art approaches
on existing small-scale datasets, while scaling to the real-world datasets we
introduce. Moreover, our model is able to `disentangle' content and sentiment
words: we automatically learn content words that are indicative of a particular
aspect as well as the aspect-specific sentiment words that are indicative of a
particular rating.
| 2,012 | Computation and Language |
Gender identity and lexical variation in social media | We present a study of the relationship between gender, linguistic style, and
social networks, using a novel corpus of 14,000 Twitter users. Prior
quantitative work on gender often treats this social variable as a female/male
binary; we argue for a more nuanced approach. By clustering Twitter users, we
find a natural decomposition of the dataset into various styles and topical
interests. Many clusters have strong gender orientations, but their use of
linguistic resources sometimes directly conflicts with the population-level
language statistics. We view these clusters as a more accurate reflection of
the multifaceted nature of gendered language styles. Previous corpus-based work
has also had little to say about individuals whose linguistic styles defy
population-level gender patterns. To identify such individuals, we train a
statistical classifier, and measure the classifier confidence for each
individual in the dataset. Examining individuals whose language does not match
the classifier's model for their gender, we find that they have social networks
that include significantly fewer same-gender social connections and that, in
general, social network homophily is correlated with the use of same-gender
language markers. Pairing computational methods and social theory thus offers a
new perspective on how gender emerges as individuals position themselves
relative to audiences, topics, and mainstream gender norms.
| 2,014 | Computation and Language |
Semantic Understanding of Professional Soccer Commentaries | This paper presents a novel approach to the problem of semantic parsing via
learning the correspondences between complex sentences and rich sets of events.
Our main intuition is that correct correspondences tend to occur more
frequently. Our model benefits from a discriminative notion of similarity to
learn the correspondence between sentence and an event and a ranking machinery
that scores the popularity of each correspondence. Our method can discover a
group of events (called macro-events) that best describes a sentence. We
evaluate our method on our novel dataset of professional soccer commentaries.
The empirical results show that our method significantly outperforms the
state-of-theart.
| 2,012 | Computation and Language |
Diffusion of Lexical Change in Social Media | Computer-mediated communication is driving fundamental changes in the nature
of written language. We investigate these changes by statistical analysis of a
dataset comprising 107 million Twitter messages (authored by 2.7 million unique
user accounts). Using a latent vector autoregressive model to aggregate across
thousands of words, we identify high-level patterns in diffusion of linguistic
change over the United States. Our model is robust to unpredictable changes in
Twitter's sampling rate, and provides a probabilistic characterization of the
relationship of macro-scale linguistic influence to a set of demographic and
geographic predictors. The results of this analysis offer support for prior
arguments that focus on geographical proximity and population size. However,
demographic similarity -- especially with regard to race -- plays an even more
central role, as cities with similar racial demographics are far more likely to
share linguistic influence. Rather than moving towards a single unified
"netspeak" dialect, language evolution in computer-mediated communication
reproduces existing fault lines in spoken American English.
| 2,014 | Computation and Language |
The origin of Mayan languages from Formosan language group of
Austronesian | Basic body-part names (BBPNs) were defined as body-part names in Swadesh
basic 200 words. Non-Mayan cognates of Mayan (MY) BBPNs were extensively
searched for, by comparing with non-MY vocabulary, including ca.1300 basic
words of 82 AN languages listed by Tryon (1985), etc. Thus found cognates (CGs)
in non-MY are listed in Table 1, as classified by language groups to which most
similar cognates (MSCs) of MY BBPNs belong. CGs of MY are classified to 23
mutually unrelated CG-items, of which 17.5 CG-items have their MSCs in
Austronesian (AN), giving its closest similarity score (CSS), CSS(AN) = 17.5,
which consists of 10.33 MSCs in Formosan, 1.83 MSCs in Western
Malayo-Polynesian (W.MP), 0.33 in Central MP, 0.0 in SHWNG, and 5.0 in Oceanic
[i.e., CSS(FORM)= 10.33, CSS(W.MP) = 1.88, ..., CSS(OC)= 5.0]. These CSSs for
language (sub)groups are also listed in the underline portion of every section
of (Section1 - Section 6) in Table 1. Chi-squar test (degree of freedom = 1)
using [Eq 1] and [Eqs.2] revealed that MSCs of MY BBPNs are distributed in
Formosan in significantly higher frequency (P < 0.001) than in other subgroups
of AN, as well as than in non-AN languages. MY is thus concluded to have been
derived from Formosan of AN. Eskimo shows some BBPN similarities to FORM and
MY.
| 2,012 | Computation and Language |
A Lightweight Stemmer for Gujarati | Gujarati is a resource poor language with almost no language processing tools
being available. In this paper we have shown an implementation of a rule based
stemmer of Gujarati. We have shown the creation of rules for stemming and the
richness in morphology that Gujarati possesses. We have also evaluated our
results by verifying it with a human expert.
| 2,012 | Computation and Language |
Design of English-Hindi Translation Memory for Efficient Translation | Developing parallel corpora is an important and a difficult activity for
Machine Translation. This requires manual annotation by Human Translators.
Translating same text again is a useless activity. There are tools available to
implement this for European Languages, but no such tool is available for Indian
Languages. In this paper we present a tool for Indian Languages which not only
provides automatic translations of the previously available translation but
also provides multiple translations, in cases where a sentence has multiple
translations, in ranked list of suggestive translations for a sentence.
Moreover this tool also lets translators have global and local saving options
of their work, so that they may share it with others, which further lightens
the task.
| 2,012 | Computation and Language |
Hidden Trends in 90 Years of Harvard Business Review | In this paper, we demonstrate and discuss results of our mining the abstracts
of the publications in Harvard Business Review between 1922 and 2012.
Techniques for computing n-grams, collocations, basic sentiment analysis, and
named-entity recognition were employed to uncover trends hidden in the
abstracts. We present findings about international relationships, sentiment in
HBR's abstracts, important international companies, influential technological
inventions, renown researchers in management theories, US presidents via
chronological analyses.
| 2,012 | Computation and Language |
Extraction of domain-specific bilingual lexicon from comparable corpora:
compositional translation and ranking | This paper proposes a method for extracting translations of morphologically
constructed terms from comparable corpora. The method is based on compositional
translation and exploits translation equivalences at the morpheme-level, which
allows for the generation of "fertile" translations (translation pairs in which
the target term has more words than the source term). Ranking methods relying
on corpus-based and translation-based features are used to select the best
candidate translation. We obtain an average precision of 91% on the Top1
candidate translation. The method was tested on two language pairs
(English-French and English-German) and with a small specialized comparable
corpora (400k words per language).
| 2,012 | Computation and Language |
Some Chances and Challenges in Applying Language Technologies to
Historical Studies in Chinese | We report applications of language technology to analyzing historical
documents in the Database for the Study of Modern Chinese Thoughts and
Literature (DSMCTL). We studied two historical issues with the reported
techniques: the conceptualization of "huaren" (Chinese people) and the attempt
to institute constitutional monarchy in the late Qing dynasty. We also discuss
research challenges for supporting sophisticated issues using our experience
with DSMCTL, the Database of Government Officials of the Republic of China, and
the Dream of the Red Chamber. Advanced techniques and tools for lexical,
syntactic, semantic, and pragmatic processing of language information, along
with more thorough data collection, are needed to strengthen the collaboration
between historians and computer scientists.
| 2,011 | Computation and Language |
Classification Analysis Of Authorship Fiction Texts in The Space Of
Semantic Fields | The use of naive Bayesian classifier (NB) and the classifier by the k nearest
neighbors (kNN) in classification semantic analysis of authors' texts of
English fiction has been analysed. The authors' works are considered in the
vector space the basis of which is formed by the frequency characteristics of
semantic fields of nouns and verbs. Highly precise classification of authors'
texts in the vector space of semantic fields indicates about the presence of
particular spheres of author's idiolect in this space which characterizes the
individual author's style.
| 2,012 | Computation and Language |
The Hangulphabet: A Descriptive Alphabet | This paper describes the Hangulphabet, a new writing system that should prove
useful in a number of contexts. Using the Hangulphabet, a user can instantly
see voicing, manner and place of articulation of any phoneme found in human
language. The Hangulphabet places consonant graphemes on a grid with the x-axis
representing the place of articulation and the y-axis representing manner of
articulation. Each individual grapheme contains radicals from both axes where
the points intersect. The top radical represents manner of articulation where
the bottom represents place of articulation. A horizontal line running through
the middle of the bottom radical represents voicing. For vowels, place of
articulation is located on a grid that represents the position of the tongue in
the mouth. This grid is similar to that of the IPA vowel chart (International
Phonetic Association, 1999). The difference with the Hangulphabet being the
trapezoid representing the vocal apparatus is on a slight tilt. Place of
articulation for a vowel is represented by a breakout figure from the grid.
This system can be used as an alternative to the International Phonetic
Alphabet (IPA) or as a complement to it. Beginning students of linguistics may
find it particularly useful. A Hangulphabet font has been created to facilitate
switching between the Hangulphabet and the IPA.
| 2,012 | Computation and Language |
The Model of Semantic Concepts Lattice For Data Mining Of Microblogs | The model of semantic concept lattice for data mining of microblogs has been
proposed in this work. It is shown that the use of this model is effective for
the semantic relations analysis and for the detection of associative rules of
key words.
| 2,012 | Computation and Language |
Optimal size, freshness and time-frame for voice search vocabulary | In this paper, we investigate how to optimize the vocabulary for a voice
search language model. The metric we optimize over is the out-of-vocabulary
(OoV) rate since it is a strong indicator of user experience. In a departure
from the usual way of measuring OoV rates, web search logs allow us to compute
the per-session OoV rate and thus estimate the percentage of users that
experience a given OoV rate. Under very conservative text normalization, we
find that a voice search vocabulary consisting of 2 to 2.5 million words
extracted from 1 week of search query data will result in an aggregate OoV rate
of 1%; at that size, the same OoV rate will also be experienced by 90% of
users. The number of words included in the vocabulary is a stable indicator of
the OoV rate. Altering the freshness of the vocabulary or the duration of the
time window over which the training data is gathered does not significantly
change the OoV rate. Surprisingly, a significantly larger vocabulary
(approximately 10 million words) is required to guarantee OoV rates below 1%
for 95% of the users.
| 2,012 | Computation and Language |
Large Scale Language Modeling in Automatic Speech Recognition | Large language models have been proven quite beneficial for a variety of
automatic speech recognition tasks in Google. We summarize results on Voice
Search and a few YouTube speech transcription tasks to highlight the impact
that one can expect from increasing both the amount of training data, and the
size of the language model estimated from such data. Depending on the task,
availability and amount of training data used, language model size and amount
of work and care put into integrating them in the lattice rescoring step we
observe reductions in word error rate between 6% and 10% relative, for systems
on a wide range of operating points between 17% and 52% word error rate.
| 2,012 | Computation and Language |
Transition-Based Dependency Parsing With Pluggable Classifiers | In principle, the design of transition-based dependency parsers makes it
possible to experiment with any general-purpose classifier without other
changes to the parsing algorithm. In practice, however, it often takes
substantial software engineering to bridge between the different
representations used by two software packages. Here we present extensions to
MaltParser that allow the drop-in use of any classifier conforming to the
interface of the Weka machine learning package, a wrapper for the TiMBL
memory-based learner to this interface, and experiments on multilingual
dependency parsing with a variety of classifiers. While earlier work had
suggested that memory-based learners might be a good choice for low-resource
parsing scenarios, we cannot support that hypothesis in this work. We observed
that support-vector machines give better parsing performance than the
memory-based learner, regardless of the size of the training set.
| 2,012 | Computation and Language |
Verbalizing Ontologies in Controlled Baltic Languages | Controlled natural languages (mostly English-based) recently have emerged as
seemingly informal supplementary means for OWL ontology authoring, if compared
to the formal notations that are used by professional knowledge engineers. In
this paper we present by examples controlled Latvian language that has been
designed to be compliant with the state of the art Attempto Controlled English.
We also discuss relation with controlled Lithuanian language that is being
designed in parallel.
| 2,010 | Computation and Language |
Detecting English Writing Styles For Non-native Speakers | Analyzing writing styles of non-native speakers is a challenging task. In
this paper, we analyze the comments written in the discussion pages of the
English Wikipedia. Using learning algorithms, we are able to detect native
speakers' writing style with an accuracy of 74%. Given the diversity of the
English Wikipedia users and the large number of languages they speak, we
measure the similarities among their native languages by comparing the
influence they have on their English writing style. Our results show that
languages known to have the same origin and development path have similar
footprint on their speakers' English writing style. To enable further studies,
the dataset we extracted from Wikipedia will be made available publicly.
| 2,012 | Computation and Language |
Dating Texts without Explicit Temporal Cues | This paper tackles temporal resolution of documents, such as determining when
a document is about or when it was written, based only on its text. We apply
techniques from information retrieval that predict dates via language models
over a discretized timeline. Unlike most previous works, we rely {\it solely}
on temporal cues implicit in the text. We consider both document-likelihood and
divergence based techniques and several smoothing methods for both of them. Our
best model predicts the mid-point of individuals' lives with a median of 22 and
mean error of 36 years for Wikipedia biographies from 3800 B.C. to the present
day. We also show that this approach works well when training on such
biographies and predicting dates both for non-biographical Wikipedia pages
about specific years (500 B.C. to 2010 A.D.) and for publication dates of short
stories (1798 to 2008). Together, our work shows that, even in absence of
temporal extraction resources, it is possible to achieve remarkable temporal
locality across a diverse set of texts.
| 2,012 | Computation and Language |
A Hindi Speech Actuated Computer Interface for Web Search | Aiming at increasing system simplicity and flexibility, an audio evoked based
system was developed by integrating simplified headphone and user-friendly
software design. This paper describes a Hindi Speech Actuated Computer
Interface for Web search (HSACIWS), which accepts spoken queries in Hindi
language and provides the search result on the screen. This system recognizes
spoken queries by large vocabulary continuous speech recognition (LVCSR),
retrieves relevant document by text retrieval, and provides the search result
on the Web by the integration of the Web and the voice systems. The LVCSR in
this system showed enough performance levels for speech with acoustic and
language models derived from a query corpus with target contents.
| 2,012 | Computation and Language |
A Principled Approach to Grammars for Controlled Natural Languages and
Predictive Editors | Controlled natural languages (CNL) with a direct mapping to formal logic have
been proposed to improve the usability of knowledge representation systems,
query interfaces, and formal specifications. Predictive editors are a popular
approach to solve the problem that CNLs are easy to read but hard to write.
Such predictive editors need to be able to "look ahead" in order to show all
possible continuations of a given unfinished sentence. Such lookahead features,
however, are difficult to implement in a satisfying way with existing grammar
frameworks, especially if the CNL supports complex nonlocal structures such as
anaphoric references. Here, methods and algorithms are presented for a new
grammar notation called Codeco, which is specifically designed for controlled
natural languages and predictive editors. A parsing approach for Codeco based
on an extended chart parsing algorithm is presented. A large subset of Attempto
Controlled English (ACE) has been represented in Codeco. Evaluation of this
grammar and the parser implementation shows that the approach is practical,
adequate and efficient.
| 2,013 | Computation and Language |
Semantic Polarity of Adjectival Predicates in Online Reviews | Web users produce more and more documents expressing opinions. Because these
have become important resources for customers and manufacturers, many have
focused on them. Opinions are often expressed through adjectives with positive
or negative semantic values. In extracting information from users' opinion in
online reviews, exact recognition of the semantic polarity of adjectives is one
of the most important requirements. Since adjectives have different semantic
orientations according to contexts, it is not satisfying to extract opinion
information without considering the semantic and lexical relations between the
adjectives and the feature nouns appropriate to a given domain. In this paper,
we present a classification of adjectives by polarity, and we analyze
adjectives that are undetermined in the absence of contexts. Our research
should be useful for accurately predicting semantic orientations of opinion
sentences, and should be taken into account before relying on an automatic
methods.
| 2,010 | Computation and Language |
A Rule-Based Approach For Aligning Japanese-Spanish Sentences From A
Comparable Corpora | The performance of a Statistical Machine Translation System (SMT) system is
proportionally directed to the quality and length of the parallel corpus it
uses. However for some pair of languages there is a considerable lack of them.
The long term goal is to construct a Japanese-Spanish parallel corpus to be
used for SMT, whereas, there are a lack of useful Japanese-Spanish parallel
Corpus. To address this problem, In this study we proposed a method for
extracting Japanese-Spanish Parallel Sentences from Wikipedia using POS tagging
and Rule-Based approach. The main focus of this approach is the syntactic
features of both languages. Human evaluation was performed over a sample and
shows promising results, in comparison with the baseline.
| 2,012 | Computation and Language |
Automating rule generation for grammar checkers | In this paper, I describe several approaches to automatic or semi-automatic
development of symbolic rules for grammar checkers from the information
contained in corpora. The rules obtained this way are an important addition to
manually-created rules that seem to dominate in rule-based checkers. However,
the manual process of creation of rules is costly, time-consuming and
error-prone. It seems therefore advisable to use machine-learning algorithms to
create the rules automatically or semi-automatically. The results obtained seem
to corroborate my initial hypothesis that symbolic machine learning algorithms
can be useful for acquiring new rules for grammar checking. It turns out,
however, that for practical uses, error corpora cannot be the sole source of
information used in grammar checking. I suggest therefore that only by using
different approaches, grammar-checkers, or more generally, computer-aided
proofreading tools, will be able to cover most frequent and severe mistakes and
avoid false alarms that seem to distract users.
| 2,012 | Computation and Language |
Two Algorithms for Finding $k$ Shortest Paths of a Weighted Pushdown
Automaton | We introduce efficient algorithms for finding the $k$ shortest paths of a
weighted pushdown automaton (WPDA), a compact representation of a weighted set
of strings with potential applications in parsing and machine translation. Both
of our algorithms are derived from the same weighted deductive logic
description of the execution of a WPDA using different search strategies.
Experimental results show our Algorithm 2 adds very little overhead vs. the
single shortest path algorithm, even with a large $k$.
| 2,013 | Computation and Language |
Using external sources of bilingual information for on-the-fly word
alignment | In this paper we present a new and simple language-independent method for
word-alignment based on the use of external sources of bilingual information
such as machine translation systems. We show that the few parameters of the
aligner can be trained on a very small corpus, which leads to results
comparable to those obtained by the state-of-the-art tool GIZA++ in terms of
precision. Regarding other metrics, such as alignment error rate or F-measure,
the parametric aligner, when trained on a very small gold-standard (450 pairs
of sentences), provides results comparable to those produced by GIZA++ when
trained on an in-domain corpus of around 10,000 pairs of sentences.
Furthermore, the results obtained indicate that the training is
domain-independent, which enables the use of the trained aligner 'on the fly'
on any new pair of sentences.
| 2,012 | Computation and Language |
The Clustering of Author's Texts of English Fiction in the Vector Space
of Semantic Fields | The clustering of text documents in the vector space of semantic fields and
in the semantic space with orthogonal basis has been analysed. It is shown that
using the vector space model with the basis of semantic fields is effective in
the cluster analysis algorithms of author's texts in English fiction. The
analysis of the author's texts distribution in cluster structure showed the
presence of the areas of semantic space that represent the author's ideolects
of individual authors. SVD factorization of the semantic fields matrix makes it
possible to reduce significantly the dimension of the semantic space in the
cluster analysis of author's texts.
| 2,012 | Computation and Language |
A Novel Feature-based Bayesian Model for Query Focused Multi-document
Summarization | Both supervised learning methods and LDA based topic model have been
successfully applied in the field of query focused multi-document
summarization. In this paper, we propose a novel supervised approach that can
incorporate rich sentence features into Bayesian topic models in a principled
way, thus taking advantages of both topic model and feature based supervised
learning methods. Experiments on TAC2008 and TAC2009 demonstrate the
effectiveness of our approach.
| 2,013 | Computation and Language |
Query-focused Multi-document Summarization: Combining a Novel Topic
Model with Graph-based Semi-supervised Learning | Graph-based semi-supervised learning has proven to be an effective approach
for query-focused multi-document summarization. The problem of previous
semi-supervised learning is that sentences are ranked without considering the
higher level information beyond sentence level. Researches on general
summarization illustrated that the addition of topic level can effectively
improve the summary quality. Inspired by previous researches, we propose a
two-layer (i.e. sentence layer and topic layer) graph-based semi-supervised
learning approach. At the same time, we propose a novel topic model which makes
full use of the dependence between sentences and words. Experimental results on
DUC and TAC data sets demonstrate the effectiveness of our proposed approach.
| 2,014 | Computation and Language |
On the complexity of learning a language: An improvement of Block's
algorithm | Language learning is thought to be a highly complex process. One of the
hurdles in learning a language is to learn the rules of syntax of the language.
Rules of syntax are often ordered in that before one rule can applied one must
apply another. It has been thought that to learn the order of n rules one must
go through all n! permutations. Thus to learn the order of 27 rules would
require 27! steps or 1.08889x10^{28} steps. This number is much greater than
the number of seconds since the beginning of the universe! In an insightful
analysis the linguist Block ([Block 86], pp. 62-63, p.238) showed that with the
assumption of transitivity this vast number of learning steps reduces to a mere
377 steps. We present a mathematical analysis of the complexity of Block's
algorithm. The algorithm has a complexity of order n^2 given n rules. In
addition, we improve Block's results exponentially, by introducing an algorithm
that has complexity of order less than n log n.
| 2,012 | Computation and Language |
Mining the Web for the Voice of the Herd to Track Stock Market Bubbles | We show that power-law analyses of financial commentaries from newspaper
web-sites can be used to identify stock market bubbles, supplementing
traditional volatility analyses. Using a four-year corpus of 17,713 online,
finance-related articles (10M+ words) from the Financial Times, the New York
Times, and the BBC, we show that week-to-week changes in power-law
distributions reflect market movements of the Dow Jones Industrial Average
(DJI), the FTSE-100, and the NIKKEI-225. Notably, the statistical regularities
in language track the 2007 stock market bubble, showing emerging structure in
the language of commentators, as progressively greater agreement arose in their
positive perceptions of the market. Furthermore, during the bubble period, a
marked divergence in positive language occurs as revealed by a Kullback-Leibler
analysis.
| 2,012 | Computation and Language |
Identifying Metaphor Hierarchies in a Corpus Analysis of Finance
Articles | Using a corpus of over 17,000 financial news reports (involving over 10M
words), we perform an analysis of the argument-distributions of the UP- and
DOWN-verbs used to describe movements of indices, stocks, and shares. Using
measures of the overlap in the argument distributions of these verbs and
k-means clustering of their distributions, we advance evidence for the proposal
that the metaphors referred to by these verbs are organised into hierarchical
structures of superordinate and subordinate groups.
| 2,011 | Computation and Language |
Identifying Metaphoric Antonyms in a Corpus Analysis of Finance Articles | Using a corpus of 17,000+ financial news reports (involving over 10M words),
we perform an analysis of the argument-distributions of the UP and DOWN verbs
used to describe movements of indices, stocks and shares. In Study 1
participants identified antonyms of these verbs in a free-response task and a
matching task from which the most commonly identified antonyms were compiled.
In Study 2, we determined whether the argument-distributions for the verbs in
these antonym-pairs were sufficiently similar to predict the most
frequently-identified antonym. Cosine similarity correlates moderately with the
proportions of antonym-pairs identified by people (r = 0.31). More
impressively, 87% of the time the most frequently-identified antonym is either
the first- or second-most similar pair in the set of alternatives. The
implications of these results for distributional approaches to determining
metaphoric knowledge are discussed.
| 2,011 | Computation and Language |
Diachronic Variation in Grammatical Relations | We present a method of finding and analyzing shifts in grammatical relations
found in diachronic corpora. Inspired by the econometric technique of measuring
return and volatility instead of relative frequencies, we propose them as a way
to better characterize changes in grammatical patterns like nominalization,
modification and comparison. To exemplify the use of these techniques, we
examine a corpus of NIPS papers and report trends which manifest at the token,
part-of-speech and grammatical levels. Building up from frequency observations
to a second-order analysis, we show that shifts in frequencies overlook deeper
trends in language, even when part-of-speech information is included. Examining
token, POS and grammatical levels of variation enables a summary view of
diachronic text as a whole. We conclude with a discussion about how these
methods can inform intuitions about specialist domains as well as changes in
language use as a whole.
| 2,012 | Computation and Language |
Language Without Words: A Pointillist Model for Natural Language
Processing | This paper explores two separate questions: Can we perform natural language
processing tasks without a lexicon?; and, Should we? Existing natural language
processing techniques are either based on words as units or use units such as
grams only for basic classification tasks. How close can a machine come to
reasoning about the meanings of words and phrases in a corpus without using any
lexicon, based only on grams?
Our own motivation for posing this question is based on our efforts to find
popular trends in words and phrases from online Chinese social media. This form
of written Chinese uses so many neologisms, creative character placements, and
combinations of writing systems that it has been dubbed the "Martian Language."
Readers must often use visual queues, audible queues from reading out loud, and
their knowledge and understanding of current events to understand a post. For
analysis of popular trends, the specific problem is that it is difficult to
build a lexicon when the invention of new ways to refer to a word or concept is
easy and common. For natural language processing in general, we argue in this
paper that new uses of language in social media will challenge machines'
abilities to operate with words as the basic unit of understanding, not only in
Chinese but potentially in other languages.
| 2,012 | Computation and Language |
Sentence Compression in Spanish driven by Discourse Segmentation and
Language Models | Previous works demonstrated that Automatic Text Summarization (ATS) by
sentences extraction may be improved using sentence compression. In this work
we present a sentence compressions approach guided by level-sentence discourse
segmentation and probabilistic language models (LM). The results presented here
show that the proposed solution is able to generate coherent summaries with
grammatical compressed sentences. The approach is simple enough to be
transposed into other languages.
| 2,012 | Computation and Language |
A comparative study of root-based and stem-based approaches for
measuring the similarity between arabic words for arabic text mining
applications | Representation of semantic information contained in the words is needed for
any Arabic Text Mining applications. More precisely, the purpose is to better
take into account the semantic dependencies between words expressed by the
co-occurrence frequencies of these words. There have been many proposals to
compute similarities between words based on their distributions in contexts. In
this paper, we compare and contrast the effect of two preprocessing techniques
applied to Arabic corpus: Rootbased (Stemming), and Stem-based (Light Stemming)
approaches for measuring the similarity between Arabic words with the well
known abstractive model -Latent Semantic Analysis (LSA)- with a wide variety of
distance functions and similarity measures, such as the Euclidean Distance,
Cosine Similarity, Jaccard Coefficient, and the Pearson Correlation
Coefficient. The obtained results show that, on the one hand, the variety of
the corpus produces more accurate results; on the other hand, the Stem-based
approach outperformed the Root-based one because this latter affects the words
meanings.
| 2,012 | Computation and Language |
Assessing Sentiment Strength in Words Prior Polarities | Many approaches to sentiment analysis rely on lexica where words are tagged
with their prior polarity - i.e. if a word out of context evokes something
positive or something negative. In particular, broad-coverage resources like
SentiWordNet provide polarities for (almost) every word. Since words can have
multiple senses, we address the problem of how to compute the prior polarity of
a word starting from the polarity of each sense and returning its polarity
strength as an index between -1 and 1. We compare 14 such formulae that appear
in the literature, and assess which one best approximates the human judgement
of prior polarities, with both regression and classification models.
| 2,012 | Computation and Language |
Natural Language Understanding Based on Semantic Relations between
Sentences | In this paper, we define event expression over sentences of natural language
and semantic relations between events. Based on this definition, we formally
consider text understanding process having events as basic unit.
| 2,012 | Computation and Language |
Good parts first - a new algorithm for approximate search in lexica and
string databases | We present a new efficient method for approximate search in electronic
lexica. Given an input string (the pattern) and a similarity threshold, the
algorithm retrieves all entries of the lexicon that are sufficiently similar to
the pattern. Search is organized in subsearches that always start with an exact
partial match where a substring of the input pattern is aligned with a
substring of a lexicon word. Afterwards this partial match is extended stepwise
to larger substrings. For aligning further parts of the pattern with
corresponding parts of lexicon entries, more errors are tolerated at each
subsequent step. For supporting this alignment order, which may start at any
part of the pattern, the lexicon is represented as a structure that enables
immediate access to any substring of a lexicon word and permits the extension
of such substrings in both directions. Experimental evaluations of the
approximate search procedure are given that show significant efficiency
improvements compared to existing techniques. Since the technique can be used
for large error bounds it offers interesting possibilities for approximate
search in special collections of "long" strings, such as phrases, sentences, or
book ti
| 2,015 | Computation and Language |
Syntactic Analysis Based on Morphological Characteristic Features of the
Romanian Language | This paper refers to the syntactic analysis of phrases in Romanian, as an
important process of natural language processing. We will suggest a real-time
solution, based on the idea of using some words or groups of words that
indicate grammatical category; and some specific endings of some parts of
sentence. Our idea is based on some characteristics of the Romanian language,
where some prepositions, adverbs or some specific endings can provide a lot of
information about the structure of a complex sentence. Such characteristics can
be found in other languages, too, such as French. Using a special grammar, we
developed a system (DIASEXP) that can perform a dialogue in natural language
with assertive and interogative sentences about a "story" (a set of sentences
describing some events from the real life).
| 2,013 | Computation and Language |
TEI and LMF crosswalks | The present paper explores various arguments in favour of making the Text
Encoding Initia-tive (TEI) guidelines an appropriate serialisation for ISO
standard 24613:2008 (LMF, Lexi-cal Mark-up Framework) . It also identifies the
issues that would have to be resolved in order to reach an appropriate
implementation of these ideas, in particular in terms of infor-mational
coverage. We show how the customisation facilities offered by the TEI
guidelines can provide an adequate background, not only to cover missing
components within the current Dictionary chapter of the TEI guidelines, but
also to allow specific lexical projects to deal with local constraints. We
expect this proposal to be a basis for a future ISO project in the context of
the on going revision of LMF.
| 2,015 | Computation and Language |
Determining token sequence mistakes in responses to questions with open
text answer | When learning grammar of the new language, a teacher should routinely check
student's exercises for grammatical correctness. The paper describes a method
of automatically detecting and reporting grammar mistakes, regarding an order
of tokens in the response. It could report extra tokens, missing tokens and
misplaced tokens. The method is useful when teaching language, where order of
tokens is important, which includes most formal languages and some natural ones
(like English). The method was implemented in a question type plug-in
CorrectWriting for the widely used learning manage system Moodle.
| 2,013 | Computation and Language |
Cutting Recursive Autoencoder Trees | Deep Learning models enjoy considerable success in Natural Language
Processing. While deep architectures produce useful representations that lead
to improvements in various tasks, they are often difficult to interpret. This
makes the analysis of learned structures particularly difficult. In this paper,
we rely on empirical tests to see whether a particular structure makes sense.
We present an analysis of the Semi-Supervised Recursive Autoencoder, a
well-known model that produces structural representations of text. We show that
for certain tasks, the structure of the autoencoder can be significantly
reduced without loss of classification accuracy and we evaluate the produced
structures using human judgment.
| 2,013 | Computation and Language |
SpeedRead: A Fast Named Entity Recognition Pipeline | Online content analysis employs algorithmic methods to identify entities in
unstructured text. Both machine learning and knowledge-base approaches lie at
the foundation of contemporary named entities extraction systems. However, the
progress in deploying these approaches on web-scale has been been hampered by
the computational cost of NLP over massive text corpora. We present SpeedRead
(SR), a named entity recognition pipeline that runs at least 10 times faster
than Stanford NLP pipeline. This pipeline consists of a high performance Penn
Treebank- compliant tokenizer, close to state-of-art part-of-speech (POS)
tagger and knowledge-based named entity recognizer.
| 2,013 | Computation and Language |
The Manifold of Human Emotions | Sentiment analysis predicts the presence of positive or negative emotions in
a text document. In this paper, we consider higher dimensional extensions of
the sentiment concept, which represent a richer set of human emotions. Our
approach goes beyond previous work in that our model contains a continuous
manifold rather than a finite set of human emotions. We investigate the
resulting model, compare it to psychological observations, and explore its
predictive capabilities.
| 2,013 | Computation and Language |
A Rhetorical Analysis Approach to Natural Language Processing | The goal of this research was to find a way to extend the capabilities of
computers through the processing of language in a more human way, and present
applications which demonstrate the power of this method. This research presents
a novel approach, Rhetorical Analysis, to solving problems in Natural Language
Processing (NLP). The main benefit of Rhetorical Analysis, as opposed to
previous approaches, is that it does not require the accumulation of large sets
of training data, but can be used to solve a multitude of problems within the
field of NLP. The NLP problems investigated with Rhetorical Analysis were the
Author Identification problem - predicting the author of a piece of text based
on its rhetorical strategies, Election Prediction - predicting the winner of a
presidential candidate's re-election campaign based on rhetorical strategies
within that president's inaugural address, Natural Language Generation - having
a computer produce text containing rhetorical strategies, and Document
Summarization. The results of this research indicate that an Author
Identification system based on Rhetorical Analysis could predict the correct
author 100% of the time, that a re-election predictor based on Rhetorical
Analysis could predict the correct winner of a re-election campaign 55% of the
time, that a Natural Language Generation system based on Rhetorical Analysis
could output text with up to 87.3% similarity to Shakespeare in style, and that
a Document Summarization system based on Rhetorical Analysis could extract
highly relevant sentences. Overall, this study demonstrated that Rhetorical
Analysis could be a useful approach to solving problems in NLP.
| 2,013 | Computation and Language |
Joint Space Neural Probabilistic Language Model for Statistical Machine
Translation | A neural probabilistic language model (NPLM) provides an idea to achieve the
better perplexity than n-gram language model and their smoothed language
models. This paper investigates application area in bilingual NLP, specifically
Statistical Machine Translation (SMT). We focus on the perspectives that NPLM
has potential to open the possibility to complement potentially `huge'
monolingual resources into the `resource-constraint' bilingual resources. We
introduce an ngram-HMM language model as NPLM using the non-parametric Bayesian
construction. In order to facilitate the application to various tasks, we
propose the joint space model of ngram-HMM language model. We show an
experiment of system combination in the area of SMT. One discovery was that our
treatment of noise improved the results 0.20 BLEU points if NPLM is trained in
relatively small corpus, in our case 500,000 sentence pairs, which is often the
case due to the long training time of NPLM.
| 2,017 | Computation and Language |
Learning New Facts From Knowledge Bases With Neural Tensor Networks and
Semantic Word Vectors | Knowledge bases provide applications with the benefit of easily accessible,
systematic relational knowledge but often suffer in practice from their
incompleteness and lack of knowledge of new entities and relations. Much work
has focused on building or extending them by finding patterns in large
unannotated text corpora. In contrast, here we mainly aim to complete a
knowledge base by predicting additional true relationships between entities,
based on generalizations that can be discerned in the given knowledgebase. We
introduce a neural tensor network (NTN) model which predicts new relationship
entries that can be added to the database. This model can be improved by
initializing entity representations with word vectors learned in an
unsupervised fashion from text, and when doing this, existing relations can
even be queried for entities that were not present in the database. Our model
generalizes and outperforms existing models for this problem, and can classify
unseen relationships in WordNet with an accuracy of 75.8%.
| 2,013 | Computation and Language |
Two SVDs produce more focal deep learning representations | A key characteristic of work on deep learning and neural networks in general
is that it relies on representations of the input that support generalization,
robust inference, domain adaptation and other desirable functionalities. Much
recent progress in the field has focused on efficient and effective methods for
computing representations. In this paper, we propose an alternative method that
is more efficient than prior work and produces representations that have a
property we call focality -- a property we hypothesize to be important for
neural network representations. The method consists of a simple application of
two consecutive SVDs and is inspired by Anandkumar (2012).
| 2,013 | Computation and Language |
Efficient Estimation of Word Representations in Vector Space | We propose two novel model architectures for computing continuous vector
representations of words from very large data sets. The quality of these
representations is measured in a word similarity task, and the results are
compared to the previously best performing techniques based on different types
of neural networks. We observe large improvements in accuracy at much lower
computational cost, i.e. it takes less than a day to learn high quality word
vectors from a 1.6 billion words data set. Furthermore, we show that these
vectors provide state-of-the-art performance on our test set for measuring
syntactic and semantic word similarities.
| 2,013 | Computation and Language |
Language learning from positive evidence, reconsidered: A
simplicity-based approach | Children learn their native language by exposure to their linguistic and
communicative environment, but apparently without requiring that their mistakes
are corrected. Such learning from positive evidence has been viewed as raising
logical problems for language acquisition. In particular, without correction,
how is the child to recover from conjecturing an over-general grammar, which
will be consistent with any sentence that the child hears? There have been many
proposals concerning how this logical problem can be dissolved. Here, we review
recent formal results showing that the learner has sufficient data to learn
successfully from positive evidence, if it favours the simplest encoding of the
linguistic input. Results include the ability to learn a linguistic prediction,
grammaticality judgements, language production, and form-meaning mappings. The
simplicity approach can also be scaled-down to analyse the ability to learn a
specific linguistic constructions, and is amenable to empirical test as a
framework for describing human language acquisition.
| 2,013 | Computation and Language |
Transfer Topic Modeling with Ease and Scalability | The increasing volume of short texts generated on social media sites, such as
Twitter or Facebook, creates a great demand for effective and efficient topic
modeling approaches. While latent Dirichlet allocation (LDA) can be applied, it
is not optimal due to its weakness in handling short texts with fast-changing
topics and scalability concerns. In this paper, we propose a transfer learning
approach that utilizes abundant labeled documents from other domains (such as
Yahoo! News or Wikipedia) to improve topic modeling, with better model fitting
and result interpretation. Specifically, we develop Transfer Hierarchical LDA
(thLDA) model, which incorporates the label information from other domains via
informative priors. In addition, we develop a parallel implementation of our
model for large-scale applications. We demonstrate the effectiveness of our
thLDA model on both a microblogging dataset and standard text collections
including AP and RCV1 datasets.
| 2,013 | Computation and Language |
Multi-Step Regression Learning for Compositional Distributional
Semantics | We present a model for compositional distributional semantics related to the
framework of Coecke et al. (2010), and emulating formal semantics by
representing functions as tensors and arguments as vectors. We introduce a new
learning method for tensors, generalising the approach of Baroni and Zamparelli
(2010). We evaluate it on two benchmark data sets, and find it to outperform
existing leading methods. We argue in our analysis that the nature of this
learning method also renders it suitable for solving more subtle problems
compositional distributional models might face.
| 2,013 | Computation and Language |
PyPLN: a Distributed Platform for Natural Language Processing | This paper presents a distributed platform for Natural Language Processing
called PyPLN. PyPLN leverages a vast array of NLP and text processing open
source tools, managing the distribution of the workload on a variety of
configurations: from a single server to a cluster of linux servers. PyPLN is
developed using Python 2.7.3 but makes it very easy to incorporate other
softwares for specific tasks as long as a linux version is available. PyPLN
facilitates analyses both at document and corpus level, simplifying management
and publication of corpora and analytical results through an easy to use web
interface. In the current (beta) release, it supports English and Portuguese
languages with support to other languages planned for future releases. To
support the Portuguese language PyPLN uses the PALAVRAS parser\citep{Bick2000}.
Currently PyPLN offers the following features: Text extraction with encoding
normalization (to UTF-8), part-of-speech tagging, token frequency, semantic
annotation, n-gram extraction, word and sentence repertoire, and full-text
search across corpora. The platform is licensed as GPL-v3.
| 2,013 | Computation and Language |
Large Scale Distributed Acoustic Modeling With Back-off N-grams | The paper revives an older approach to acoustic modeling that borrows from
n-gram language modeling in an attempt to scale up both the amount of training
data and model size (as measured by the number of parameters in the model), to
approximately 100 times larger than current sizes used in automatic speech
recognition. In such a data-rich setting, we can expand the phonetic context
significantly beyond triphones, as well as increase the number of Gaussian
mixture components for the context-dependent states that allow it. We have
experimented with contexts that span seven or more context-independent phones,
and up to 620 mixture components per state. Dealing with unseen phonetic
contexts is accomplished using the familiar back-off technique used in language
modeling due to implementation simplicity. The back-off acoustic model is
estimated, stored and served using MapReduce distributed computing
infrastructure.
Speech recognition experiments are carried out in an N-best list rescoring
framework for Google Voice Search. Training big models on large amounts of data
proves to be an effective way to increase the accuracy of a state-of-the-art
automatic speech recognition system. We use 87,000 hours of training data
(speech along with transcription) obtained by filtering utterances in Voice
Search logs on automatic speech recognition confidence. Models ranging in size
between 20--40 million Gaussians are estimated using maximum likelihood
training. They achieve relative reductions in word-error-rate of 11% and 6%
when combined with first-pass models trained using maximum likelihood, and
boosted maximum mutual information, respectively. Increasing the context size
beyond five phones (quinphones) does not help.
| 2,013 | Computation and Language |
Towards the Rapid Development of a Natural Language Understanding Module | When developing a conversational agent, there is often an urgent need to have
a prototype available in order to test the application with real users. A
Wizard of Oz is a possibility, but sometimes the agent should be simply
deployed in the environment where it will be used. Here, the agent should be
able to capture as many interactions as possible and to understand how people
react to failure. In this paper, we focus on the rapid development of a natural
language understanding module by non experts. Our approach follows the learning
paradigm and sees the process of understanding natural language as a
classification problem. We test our module with a conversational agent that
answers questions in the art domain. Moreover, we show how our approach can be
used by a natural language interface to a cinema database.
| 2,011 | Computation and Language |
S\'emantique des d\'eterminants dans un cadre richement typ\'e | The variation of word meaning according to the context leads us to enrich the
type system of our syntactical and semantic analyser of French based on
categorial grammars and Montague semantics (or lambda-DRT). The main advantage
of a deep semantic analyse is too represent meaning by logical formulae that
can be easily used e.g. for inferences. Determiners and quantifiers play a
fundamental role in the construction of those formulae. But in our rich type
system the usual semantic terms do not work. We propose a solution ins- pired
by the tau and epsilon operators of Hilbert, kinds of generic elements and
choice functions. This approach unifies the treatment of the different determi-
ners and quantifiers as well as the dynamic binding of pronouns. Above all,
this fully computational view fits in well within the wide coverage parser
Grail, both from a theoretical and a practical viewpoint.
| 2,013 | Computation and Language |
Lexical Access for Speech Understanding using Minimum Message Length
Encoding | The Lexical Access Problem consists of determining the intended sequence of
words corresponding to an input sequence of phonemes (basic speech sounds) that
come from a low-level phoneme recognizer. In this paper we present an
information-theoretic approach based on the Minimum Message Length Criterion
for solving the Lexical Access Problem. We model sentences using phoneme
realizations seen in training, and word and part-of-speech information obtained
from text corpora. We show results on multiple-speaker, continuous, read speech
and discuss a heuristic using equivalence classes of similar sounding words
which speeds up the recognition process without significant deterioration in
recognition accuracy.
| 2,013 | Computation and Language |
Building a reordering system using tree-to-string hierarchical model | This paper describes our submission to the First Workshop on Reordering for
Statistical Machine Translation. We have decided to build a reordering system
based on tree-to-string model, using only publicly available tools to
accomplish this task. With the provided training data we have built a
translation model using Moses toolkit, and then we applied a chart decoder,
implemented in Moses, to reorder the sentences. Even though our submission only
covered English-Farsi language pair, we believe that the approach itself should
work regardless of the choice of the languages, so we have also carried out the
experiments for English-Italian and English-Urdu. For these language pairs we
have noticed a significant improvement over the baseline in BLEU, Kendall-Tau
and Hamming metrics. A detailed description is given, so that everyone can
reproduce our results. Also, some possible directions for further improvements
are discussed.
| 2,013 | Computation and Language |
Termhood-based Comparability Metrics of Comparable Corpus in Special
Domain | Cross-Language Information Retrieval (CLIR) and machine translation (MT)
resources, such as dictionaries and parallel corpora, are scarce and hard to
come by for special domains. Besides, these resources are just limited to a few
languages, such as English, French, and Spanish and so on. So, obtaining
comparable corpora automatically for such domains could be an answer to this
problem effectively. Comparable corpora, that the subcorpora are not
translations of each other, can be easily obtained from web. Therefore,
building and using comparable corpora is often a more feasible option in
multilingual information processing. Comparability metrics is one of key issues
in the field of building and using comparable corpus. Currently, there is no
widely accepted definition or metrics method of corpus comparability. In fact,
Different definitions or metrics methods of comparability might be given to
suit various tasks about natural language processing. A new comparability,
namely, termhood-based metrics, oriented to the task of bilingual terminology
extraction, is proposed in this paper. In this method, words are ranked by
termhood not frequency, and then the cosine similarities, calculated based on
the ranking lists of word termhood, is used as comparability. Experiments
results show that termhood-based metrics performs better than traditional
frequency-based metrics.
| 2,013 | Computation and Language |
Bilingual Terminology Extraction Using Multi-level Termhood | Purpose: Terminology is the set of technical words or expressions used in
specific contexts, which denotes the core concept in a formal discipline and is
usually applied in the fields of machine translation, information retrieval,
information extraction and text categorization, etc. Bilingual terminology
extraction plays an important role in the application of bilingual dictionary
compilation, bilingual Ontology construction, machine translation and
cross-language information retrieval etc. This paper addresses the issues of
monolingual terminology extraction and bilingual term alignment based on
multi-level termhood.
Design/methodology/approach: A method based on multi-level termhood is
proposed. The new method computes the termhood of the terminology candidate as
well as the sentence that includes the terminology by the comparison of the
corpus. Since terminologies and general words usually have differently
distribution in the corpus, termhood can also be used to constrain and enhance
the performance of term alignment when aligning bilingual terms on the parallel
corpus. In this paper, bilingual term alignment based on termhood constraints
is presented.
Findings: Experiment results show multi-level termhood can get better
performance than existing method for terminology extraction. If termhood is
used as constrain factor, the performance of bilingual term alignment can be
improved.
| 2,012 | Computation and Language |
Compactified Horizontal Visibility Graph for the Language Network | A compactified horizontal visibility graph for the language network is
proposed. It was found that the networks constructed in such way are scale
free, and have a property that among the nodes with largest degrees there are
words that determine not only a text structure communication, but also its
informational structure.
| 2,013 | Computation and Language |
Towards a Semantic-based Approach for Modeling Regulatory Documents in
Building Industry | Regulations in the Building Industry are becoming increasingly complex and
involve more than one technical area. They cover products, components and
project implementation. They also play an important role to ensure the quality
of a building, and to minimize its environmental impact. In this paper, we are
particularly interested in the modeling of the regulatory constraints derived
from the Technical Guides issued by CSTB and used to validate Technical
Assessments. We first describe our approach for modeling regulatory constraints
in the SBVR language, and formalizing them in the SPARQL language. Second, we
describe how we model the processes of compliance checking described in the
CSTB Technical Guides. Third, we show how we implement these processes to
assist industrials in drafting Technical Documents in order to acquire a
Technical Assessment; a compliance report is automatically generated to explain
the compliance or noncompliance of this Technical Documents.
| 2,012 | Computation and Language |
Probabilistic Frame Induction | In natural-language discourse, related events tend to appear near each other
to describe a larger scenario. Such structures can be formalized by the notion
of a frame (a.k.a. template), which comprises a set of related events and
prototypical participants and event transitions. Identifying frames is a
prerequisite for information extraction and natural language generation, and is
usually done manually. Methods for inducing frames have been proposed recently,
but they typically use ad hoc procedures and are difficult to diagnose or
extend. In this paper, we propose the first probabilistic approach to frame
induction, which incorporates frames, events, participants as latent topics and
learns those frame and event transitions that best explain the text. The number
of frames is inferred by a novel application of a split-merge method from
syntactic parsing. In end-to-end evaluations from text to induced frames and
extracted facts, our method produced state-of-the-art results while
substantially reducing engineering effort.
| 2,013 | Computation and Language |
NLP and CALL: integration is working | In the first part of this article, we explore the background of
computer-assisted learning from its beginnings in the early XIXth century and
the first teaching machines, founded on theories of learning, at the start of
the XXth century. With the arrival of the computer, it became possible to offer
language learners different types of language activities such as comprehension
tasks, simulations, etc. However, these have limits that cannot be overcome
without some contribution from the field of natural language processing (NLP).
In what follows, we examine the challenges faced and the issues raised by
integrating NLP into CALL. We hope to demonstrate that the key to success in
integrating NLP into CALL is to be found in multidisciplinary work between
computer experts, linguists, language teachers, didacticians and NLP
specialists.
| 2,006 | Computation and Language |
A Labeled Graph Kernel for Relationship Extraction | In this paper, we propose an approach for Relationship Extraction (RE) based
on labeled graph kernels. The kernel we propose is a particularization of a
random walk kernel that exploits two properties previously studied in the RE
literature: (i) the words between the candidate entities or connecting them in
a syntactic representation are particularly likely to carry information
regarding the relationship; and (ii) combining information from distinct
sources in a kernel may help the RE system make better decisions. We performed
experiments on a dataset of protein-protein interactions and the results show
that our approach obtains effectiveness values that are comparable with the
state-of-the art kernel methods. Moreover, our approach is able to outperform
the state-of-the-art kernels when combined with other kernel methods.
| 2,013 | Computation and Language |
Role of temporal inference in the recognition of textual inference | This project is a part of nature language processing and its aims to develop
a system of recognition inference text-appointed TIMINF. This type of system
can detect, given two portions of text, if a text is semantically deducted from
the other. We focused on making the inference time in this type of system. For
that we have built and analyzed a body built from questions collected through
the web. This study has enabled us to classify different types of times
inferences and for designing the architecture of TIMINF which seeks to
integrate a module inference time in a detection system inference text. We also
assess the performance of sorties TIMINF system on a test corpus with the same
strategy adopted in the challenge RTE.
| 2,013 | Computation and Language |
Development of Yes/No Arabic Question Answering System | Developing Question Answering systems has been one of the important research
issues because it requires insights from a variety of
disciplines,including,Artificial Intelligence,Information Retrieval,
Information Extraction,Natural Language Processing, and Psychology.In this
paper we realize a formal model for a lightweight semantic based open domain
yes/no Arabic question answering system based on paragraph retrieval with
variable length. We propose a constrained semantic representation. Using an
explicit unification framework based on semantic similarities and query
expansion synonyms and antonyms.This frequently improves the precision of the
system. Employing the passage retrieval system achieves a better precision by
retrieving more paragraphs that contain relevant answers to the question; It
significantly reduces the amount of text to be processed by the system.
| 2,013 | Computation and Language |
Non-simplifying Graph Rewriting Termination | So far, a very large amount of work in Natural Language Processing (NLP) rely
on trees as the core mathematical structure to represent linguistic
informations (e.g. in Chomsky's work). However, some linguistic phenomena do
not cope properly with trees. In a former paper, we showed the benefit of
encoding linguistic structures by graphs and of using graph rewriting rules to
compute on those structures. Justified by some linguistic considerations, graph
rewriting is characterized by two features: first, there is no node creation
along computations and second, there are non-local edge modifications. Under
these hypotheses, we show that uniform termination is undecidable and that
non-uniform termination is decidable. We describe two termination techniques
based on weights and we give complexity bound on the derivation length for
these rewriting system.
| 2,013 | Computation and Language |
Ending-based Strategies for Part-of-speech Tagging | Probabilistic approaches to part-of-speech tagging rely primarily on
whole-word statistics about word/tag combinations as well as contextual
information. But experience shows about 4 per cent of tokens encountered in
test sets are unknown even when the training set is as large as a million
words. Unseen words are tagged using secondary strategies that exploit word
features such as endings, capitalizations and punctuation marks. In this work,
word-ending statistics are primary and whole-word statistics are secondary.
First, a tagger was trained and tested on word endings only. Subsequent
experiments added back whole-word statistics for the words occurring most
frequently in the training set. As grew larger, performance was expected to
improve, in the limit performing the same as word-based taggers. Surprisingly,
the ending-based tagger initially performed nearly as well as the word-based
tagger; in the best case, its performance significantly exceeded that of the
word-based tagger. Lastly, and unexpectedly, an effect of negative returns was
observed - as grew larger, performance generally improved and then declined. By
varying factors such as ending length and tag-list strategy, we achieved a
success rate of 97.5 percent.
| 2,013 | Computation and Language |
KSU KDD: Word Sense Induction by Clustering in Topic Space | We describe our language-independent unsupervised word sense induction
system. This system only uses topic features to cluster different word senses
in their global context topic space. Using unlabeled data, this system trains a
latent Dirichlet allocation (LDA) topic model then uses it to infer the topics
distribution of the test instances. By clustering these topics distributions in
their topic space we cluster them into different senses. Our hypothesis is that
closeness in topic space reflects similarity between different word senses.
This system participated in SemEval-2 word sense induction and disambiguation
task and achieved the second highest V-measure score among all other systems.
| 2,010 | Computation and Language |
Structure-semantics interplay in complex networks and its effects on the
predictability of similarity in texts | There are different ways to define similarity for grouping similar texts into
clusters, as the concept of similarity may depend on the purpose of the task.
For instance, in topic extraction similar texts mean those within the same
semantic field, whereas in author recognition stylistic features should be
considered. In this study, we introduce ways to classify texts employing
concepts of complex networks, which may be able to capture syntactic, semantic
and even pragmatic features. The interplay between the various metrics of the
complex networks is analyzed with three applications, namely identification of
machine translation (MT) systems, evaluation of quality of machine translated
texts and authorship recognition. We shall show that topological features of
the networks representing texts can enhance the ability to identify MT systems
in particular cases. For evaluating the quality of MT texts, on the other hand,
high correlation was obtained with methods capable of capturing the semantics.
This was expected because the golden standards used are themselves based on
word co-occurrence. Notwithstanding, the Katz similarity, which involves
semantic and structure in the comparison of texts, achieved the highest
correlation with the NIST measurement, indicating that in some cases the
combination of both approaches can improve the ability to quantify quality in
MT. In authorship recognition, again the topological features were relevant in
some contexts, though for the books and authors analyzed good results were
obtained with semantic features as well. Because hybrid approaches encompassing
semantic and topological features have not been extensively used, we believe
that the methodology proposed here may be useful to enhance text classification
considerably, as it combines well-established strategies.
| 2,012 | Computation and Language |
Statistical sentiment analysis performance in Opinum | The classification of opinion texts in positive and negative is becoming a
subject of great interest in sentiment analysis. The existence of many labeled
opinions motivates the use of statistical and machine-learning methods.
First-order statistics have proven to be very limited in this field. The Opinum
approach is based on the order of the words without using any syntactic and
semantic information. It consists of building one probabilistic model for the
positive and another one for the negative opinions. Then the test opinions are
compared to both models and a decision and confidence measure are calculated.
In order to reduce the complexity of the training corpus we first lemmatize the
texts and we replace most named-entities with wildcards. Opinum presents an
accuracy above 81% for Spanish opinions in the financial products domain. In
this work we discuss which are the most important factors that have impact on
the classification performance.
| 2,013 | Computation and Language |