Pasar al contenido principal

TECNOLOGIAS DE LA LENGUA PARA ANALISIS DE OPINIONES EN REDES SOCIALES

FFI2014-51978-C2-1-R

•
Nombre agencia financiadora Ministerio de Economía y Competitividad
Acrónimo agencia financiadora MINECO
Programa Programa Estatal de I+D+i Orientado a los Retos de la Sociedad
Convocatoria Proyectos I+D+i Retos de Investigación
Año convocatoria 2014
Plan Estatal 2013-2016 Plan Estatal de Investigación Científica y Técnica y de Innovación
Unidad de gestión Secretaría de Estado de Educación, Formación Profesional y Universidades
Centro beneficiario UNIVERSIDAD DE VIGO
Identificador persistente http://dx.doi.org/10.13039/501100003329

Resultados relacionados

Resultados totales (Incluyendo duplicados): 26
Encontrada(s) 1 página(s)

Minerva. Repositorio Institucional de la Universidad de Santiago de Compostela
Publicación Artículo científico (JournalArticle)

LinguaKit: uma ferramenta multilingue para a analise linguistica e a extraçao de informaçao, LinguaKit: a multilingual tool for linguistic analysis and information extraction

  • Gamallo Otero, Pablo
  • García González, Marcos
Este artigo apresenta LinguaKit, uma suite multilingue de ferramentas de analise, extraçao, anotaçao e correçao linguısticas. LinguaKit permite realizar tarefas tao diversas como a lematizaçao, a etiquetagem morfossintatica ou a analise sintatica (entre outras), incluindo tambem aplicaçoes para a analise de sentimentos (ou minaria de opinioes), a extraçao de termos multipalavra, ou a anotaçao concetual e ligaçao a recursos enciclopedicos tais como a DBpedia. A maior parte dos modulos funcionam para quatro variedades linguısticas: portugues, espanhol, ingles e galego. Alinguagem de programaçao de LinguaKit ́e Perl, e o codigo esta disponıvel sob a licença livre GPLv3
Minerva. Repositorio Institucional de la Universidad de Santiago de Compostela
Publicación Artículo científico (JournalArticle)

The role of syntactic dependencies in compositional distributional semantics

  • Gamallo Otero, Pablo
This article provides a preliminary semantic framework for Dependency Grammar in which lexical words are semantically defined as contextual distributions (sets of contexts) while syntactic dependencies are compositional operations on word distributions. More precisely, any syntactic dependency uses the contextual distribution of the dependent word to restrict the distribution of the head, and makes use of the contextual distribution of the head to restrict that of the dependent word. The interpretation of composite expressions and sentences, which are analyzed as a tree of binary dependencies, is performed by restricting the contexts of words dependency by dependency in a left-to-right incremental way. Consequently, the meaning of the whole composite expression or sentence is not a single representation, but a list of contextualized senses, namely the restricted distributions of its constituent (lexical) words. We report the results of two large-scale corpus-based experiments on two different natural language processing applications: paraphrasing and compositional translation
Minerva. Repositorio Institucional de la Universidad de Santiago de Compostela
Publicación Artículo científico (JournalArticle)

Comparing Supervised Machine Learning Strategies and Linguistic Features to Search for Very Negative Opinions

  • Al-Matarneh Mohammad Ata, Sattam
  • Gamallo Otero, Pablo
In this paper, we examine the performance of several classifiers in the process of searching for very negative opinions. More precisely, we do an empirical study that analyzes the influence of three types of linguistic features (n-grams, word embeddings, and polarity lexicons) and their combinations when they are used to feed different supervised machine learning classifiers: Naive Bayes (NB), Decision Tree (DT), and Support Vector Machine (SVM). The experiments we have carried out show that SVM clearly outperforms NB and DT in all datasets by taking into account all features individually as well as their combinations
Minerva. Repositorio Institucional de la Universidad de Santiago de Compostela
Publicación Capítulo o parte de un libro (BookPart)

Comparing two Basic Methods for Discriminating Between Similar Languages and Varieties

  • Gamallo Otero, Pablo
  • Pichel, José Ramom
  • Alegria Loinaz, Iñaki
  • Agirrezabal, Manex
This article describes the systems submitted by the Citius Ixa Imaxin team to the Discriminating Similar Languages Shared Task 2016. The systems are based on two different strategies: classification with ranked dictionaries and Naive Bayes classifiers. The results of the evaluation show that ranking dictionaries are more sound and stable across different domains while basic bayesian models perform reasonably well on in-domain datasets, but their performance drops when they are applied on out-of-domain texts.
Minerva. Repositorio Institucional de la Universidad de Santiago de Compostela
Publicación Artículo científico (JournalArticle)

Dependency parsing with finite state transducers and compression rules

  • Gamallo Otero, Pablo
  • García González, Marcos
This article proposes a syntactic parsing strategy based on a dependency grammar containing formal rules and a compression technique that reduces the complexity of those rules. Compression parsing is mainly driven by the ‘single-head’ constraint of Dependency Grammar, and can be seen as an alternative method to the well-known constructive strategy. The compression algorithm simplifies the input sentence by progressively removing from it the dependent tokens as soon as binary syntactic dependencies are recognized. This strategy is thus similar to that used in deterministic dependency parsing. A compression parser was implemented and released under General Public License, as well as a cross-lingual grammar with Universal Dependencies, containing only broad-coverage rules applied to Romance languages. The system is an almost delexicalized parser which does not need training data to analyze Romance languages. The rule-based cross-lingual parser was submitted to CoNLL 2017 Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies. The performance of our system was compared to the other supervised systems participating in the competition, paying special attention to the parsing of different treebanks of the same language. We also trained a supervised delexicalized parser for Romance languages in order to compare it to our rule-based system. The results show that the performance of our cross-lingual method does not change across related languages and across different treebanks, while most supervised methods turn out to be very dependent on the text domain used to train the system
Minerva. Repositorio Institucional de la Universidad de Santiago de Compostela
Publicación Artículo científico (JournalArticle)

Explorando métodos non-supervisados para calcular a similitude semántica textual, Exploring Unsupervised Methods to Sematic Textual Similarity

  • Gamallo Otero, Pablo
  • Pereira Fariña, Martín
Neste traballo preséntanse varios métodos non-supervisados para a detección da similitude semántica textual, os cales están baseados en modelos distribucionais e no parseado de dependencias. Os sistemas son avaliados mediante datasets empregados na ASSIN Shared Task, celebrada conxuntamente co PROPOR 2016. Os métodos máis básicos ofrecen un mellor comportamento que aqueles, mais complexos, que inclúen información sintáctico-semántica na análise das oracións. Por último, o uso de modelos distribucionais construidos automaticamente a partir de corpus ofrece resultados comparábeis ás estratexias que utilizan recursos léxicos externos construídos manualmente.
Minerva. Repositorio Institucional de la Universidad de Santiago de Compostela
Publicación Artículo científico (JournalArticle)

LinguaKit: a Big Data-based multilingual tool for linguistic analysis and information extraction

  • Gamallo Otero, Pablo
  • García González, Marcos
  • Piñeiro Pomar, César Alfredo
  • Martínez-Castaño, Rodrigo
  • Pichel Campos, Juan Carlos
This paper presents LinguaKit, a multilingual suite of tools for analysis, extraction, annotation and linguistic correction, as well as its integration into a Big Data infrastructure. LinguaKit allows the user to perform different tasks such as PoS-tagging, syntactic parsing, coreference resolution (among others), including applications for relation extraction, sentiment analysis, summarization, extraction of multiword expressions, or entity linking to DBpedia. Most modules work in four languages: Portuguese, Spanish, English, and Galician. The system is programmed in Perl and is freely available under a GPLv3 license.
Minerva. Repositorio Institucional de la Universidad de Santiago de Compostela
Publicación Artículo científico (JournalArticle)

Using the Outlier Detection Task to Evaluate Distributional Semantic Models

  • Gamallo Otero, Pablo
In this article, we define the outlier detection task and use it to compare neural-based word embeddings with transparent count-based distributional representations. Using the English Wikipedia as a text source to train the models, we observed that embeddings outperform count-based representations when their contexts are made up of bag-of-words. However, there are no sharp differences between the two models if the word contexts are defined as syntactic dependencies. In general, syntax-based models tend to perform better than those based on bag-of-words for this specific task. Similar experiments were carried out for Portuguese with similar results. The test datasets we have created for the outlier detection task in English and Portuguese are freely available
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Modeling of learning curves with applications to POS tagging

  • Vilares Ferro, Manuel
  • Darriba Bilbao, Víctor Manuel
  • Ribadas Pena, Francisco José
An algorithm to estimate the evolution of learning curves on the whole of a training data base, based on the results obtained from a portion and using a functional strategy, is introduced. We approximate iteratively the sought value at the desired time, independently of the learning technique used and once a point in the process, called prediction level, has been passed. The proposal proves to be formally correct with respect to our working hypotheses and includes a reliable proximity condition. This allows the user to fix a convergence threshold with respect to the accuracy finally achievable, which extends the concept of stopping criterion and seems to be effective even in the presence of distorting observations. Our aim is to evaluate the training effort, supporting decision making in order to reduce the need for both human and computational resources during the learning process. The proposal is of interest in at least three operational procedures. The first is the anticipation of accuracy gain, with the purpose of measuring how much work is needed to achieve a certain degree of performance. The second relates the comparison of efficiency between systems at training time, with the objective of completing this task only for the one that best suits our requirements. The prediction of accuracy is also a valuable item of information for customizing systems, since we can estimate in advance the impact of settings on both the performance and the development costs. Using the generation of part-of-speech taggers as an example application, the experimental results are consistent with our expectations., Ministerio de Economía y Competitividad | Ref. FFI2014-51978-C2-1-R
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Diseño e implementación de una base de conocimiento terminológico sobre enfermedades raras, Design and implementation of a terminological knowledge base on rare diseases

  • Varela Vila, Tamara
  • Sánchez Trigo, Maria Elena
Este trabajo se centra en la elaboración de recursos terminográficos de base conceptual para la gestión de la terminología médica. Se presenta una base de conocimiento terminológico bilingüe (francés y español) sobre un numeroso grupo de enfermedades raras: los errores innatos del metabolismo. El recurso diseñado permite acceder a la terminología utilizada para denominar estas enfermedades, sus síntomas y signos de una forma conceptualmente estructurada. En primer lugar, tras el marco teórico, se describen las fases seguidas para la elaboración de la base de conocimiento: compilación, etiquetado y explotación del corpus de trabajo y construcción de la ontología subyacente. En segundo lugar, se recogen sus datos más significativos y se muestran las funcionalidades de la interfaz de consulta. Su finalidad es describir el subdominio y servir como fuente de información en la traducción especializada., This article focuses on the creation of conceptually-based terminographical resources for managing medical terminology. We present a bilingual (French and Spanish) terminological knowledge base on a large group of rare diseases, namely inborn errors of metabolism. This resource gives its users conceptually structured access to the terminology used when referring to these diseases and their signs and symptoms. After introducing the theoretical framework underpinning its design, we describe the various stages in the creation of the knowledge base itself: compilation, tagging, exploitation of the working corpus and construction of the underlying ontology. We then give an overview of the salient data and provide a guide to the functionalities of the query interface. The purpose of the resource is to describe the sub-domain in question and serve as a source of information for specialist translators., Ministerio de Economía y Competitividad | Ref. FFI2014-51978-C2-1-R, Agencia Estatal de Investigación | Ref. TIN2017-85160-C2-2-R
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Contribución de congreso (ConferenceOutput)

An efficient dynamic oracle for unrestricted non-projective parsing

  • Gómez Rodríguez, Carlos
  • Fernández González, Daniel
We define a dynamic oracle for the Covington non-projective dependency parser. This is not only the first dynamic oracle that supports arbitrary non-projectivity, but also considerably more efficient (O(n)) than the only existing oracle with restricted non-projectivity support. Experiments show that training with the dynamic oracle significantly improves parsing accuracy over the static oracle baseline on a wide range of treebanks., Ministerio de Economía y Competitividad | Ref. FFI2014-51978-C2-1-R, Ministerio de Economía y Competitividad | Ref. FFI2014-51978-C2-2-R, Xunta de Galicia | Ref. R2014/034
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Meemi: a simple method for post-processing and integrating cross-lingual word embeddings

  • Doval Mosquera, Yerai
  • Camacho Collados, Jose
  • Espinosa Anke, Luis
  • Schockaert, Steven
Word embeddings have become a standard resource in the toolset of any Natural Language Processing practitioner. While monolingual word embeddings encode information about words in the context of a particular language, cross-lingual embeddings define a multilingual space where word embeddings from two or more languages are integrated together. Current state-of-the-art approaches learn these embeddings by aligning two disjoint monolingual vector spaces through an orthogonal transformation which preserves the structure of the monolingual counterparts. In this work, we propose to apply an additional transformation after this initial alignment step, which aims to bring the vector representations of a given word and its translations closer to their average. Since this additional transformation is non-orthogonal, it also affects the structure of the monolingual spaces. We show that our approach both improves the integration of the monolingual spaces and the quality of the monolingual spaces themselves. Furthermore, because our transformation can be applied to an arbitrary number of languages, we are able to effectively obtain a truly multilingual space. The resulting (monolingual and multilingual) spaces show consistent gains over the current state-of-the-art in standard intrinsic tasks, namely dictionary induction and word similarity, as well as in extrinsic tasks such as cross-lingual hypernym discovery and cross-lingual natural language inference., Agencia Estatal de Investigación | Ref. TIN2017-85160-C2-2-R, Ministerio de Economía y Competitividad | Ref. FFI2014-51978-C2-1-R, Ministerio de Economía y Competitividad | Ref. BES-2015-073768, Xunta de Galicia | Ref. ED431D 2017/12, Universidade de Vigo/CISUG
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

On the performance of phonetic algorithms in microtext normalization

  • Doval Mosquera, Yerai
  • Vilares Ferro, Manuel
  • Vilares Ferro, Jesús
User–generated content published on microblogging social networks constitutes a priceless source of information. However, microtexts usually deviate from the standard lexical and grammatical rules of the language, thus making its processing by traditional intelligent systems very difficult. As an answer, microtext normalization consists in transforming those non–standard microtexts into standard well–written texts as a preprocessing step, allowing traditional approaches to continue with their usual processing. Given the importance of phonetic phenomena in non–standard text formation, an essential element of the knowledge base of a normalizer would be the phonetic rules that encode these phenomena, which can be found in the so–called phonetic algorithms. In this work we experiment with a wide range of phonetic algorithms for the English language. The aim of this study is to determine the best phonetic algorithms within the context of candidate generation for microtext normalization. In other words, we intend to find those algorithms that taking as input non–standard terms to be normalized allow us to obtain as output the smallest possible sets of normalization candidates which still contain the corresponding target standard words. As it will be stated, the choice of the phonetic algorithm will depend heavily on the capabilities of the candidate selection mechanism which we usually find at the end of a microtext normalization pipeline. The faster it can make the right choices among big enough sets of candidates, the more we can sacrifice on the precision of the phonetic algorithms in favour of coverage in order to increase the overall performance of the normalization system, Agencia Estatal de Investigación | Ref. TIN2017-85160-C2-1-R, Agencia Estatal de Investigación | Ref. TIN2017-85160-C2-2-R, Ministerio de Economía y Competitividad | Ref. FFI2014-51978-C2-1-R, Ministerio de Economía y Competitividad | Ref. FFI2014-51978-C2-2-R, Xunta de Galicia | Ref. ED431D-2017/12, Xunta de Galicia | Ref. ED431B2017/01, Xunta de Galicia | Ref. ED431D R2016/046, Ministerio de Economía y Competitividad | Ref. BES-2015-073768
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Análisis terminológico de la categoría conceptual ‘enfermedad’ en artículos de investigación originales sobre enfermedades raras en francés y español, A terminological analysis of the conceptual category ‘disease’ in original research articles on rare diseases in French and Spanish

  • Varela Vila, Tamara
  • Sánchez Trigo, Maria Elena
The aim of this article is to present the most common terms and syntactic patterns with regard to the conceptual category ‘disease’ in original research articles on rare diseases in French and Spanish. The research is specifically focused on the study of the terminology used when referring to the generic concept of health loss or health alterations. Firstly, the characteristics of the comparable corpus we have designed are presented. Secondly, the extracted terms and patterns are classified and analyzed. The results obtained show the presence within the texts of a significant number of terms which are synonymous in nature but whose forms and frequencies of usage do not always correspond in the languages considered in this research., Cet article a pour objectif d’identifier les termes et patrons syntaxiques les plus fréquents associés à la catégorie conceptuelle ‘maladie’ dans des articles originaux de recherche sur les maladies rares en français et en espagnol. Le travail vise spécifiquement l’étude de la terminologie employée pour désigner le concept générique de perte ou d’altération de l’état de santé. Dans un premier temps, nous présentons les caractéristiques du corpus comparable que nous avons créé. Dans un deuxième temps, les termes et les patrons extraits du corpus sont classés et analysés. Les résultats montrent la présence dans les textes d’un nombre important de dénominations de nature synonymique avec des formes et des fréquences d’emploi qui ne sont pas toujours concordantes dans les langues étudiées, Ministerio de Economía y Competitividad | Ref. FFI2014-51978-C2-1-R
RUA. Repositorio Institucional de la Universidad de Alicante
Publicación Otros (Other)

Segmentación de palabras en español mediante modelos del lenguaje basados en redes neuronales, Spanish word segmentation through neural language models

  • Doval, Yerai
  • Gómez Rodríguez, Carlos
  • Vilares Ferro, Jesús
En las plataformas de microblogging abundan ciertos tokens especiales como los hashtags o las menciones en los que un grupo de palabras se escriben juntas sin espaciado entre ellas; p.ej.: #añobisiesto o @ryanreynoldsnet. Debido a la forma en que se escriben este tipo de textos, este fenómeno de ensamblado de palabras puede aparecer junto a su opuesto, la segmentación de palabras, afectando a cualquier elemento del texto y dificultando su análisis. En este trabajo se muestra un enfoque algorítmico que utiliza como base un modelo del lenguaje - en nuestro caso concreto uno basado en redes neuronales - para resolver el problema de la segmentación y ensamblado de palabras, en el que se trata de recuperar el espaciado estándar de las palabras que han sufrido alguna de estas transformaciones añadiendo o quitando espacios donde corresponda. Los resultados obtenidos son prometedores e indican que tras un mayor refinamiento del modelo del lenguaje se podrá sobrepasar al estado del arte.
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

EN-ES-CS: An English-Spanish Code-Switching Twitter Corpus for Multilingual Sentiment Analysis

  • Vilares, David
  • Alonso, Miguel A.
  • Gómez-Rodríguez, Carlos
[Abstract]: Code-switching texts are those that contain terms in two or more different languages, and they appear increasingly often in social media. The aim of this paper is to provide a resource to the research community to evaluate the performance of sentiment classification techniques on this complex multilingual environment, proposing an English-Spanish corpus of tweets with code-switching (EN-ES-CS CORPUS). The tweets are labeled according to two well-known criteria used for this purpose: SentiStrength and a trinary scale (positive, neutral and negative categories). Preliminary work on the resource is already done, providing a set of baselines for the research community., This research is supported by the Ministerio de Economía y Competitividad (FFI2014-51978-C2) and Xunta de Galicia (R2014/034). David Vilares is funded by the Ministerio de Educación, Cultura y Deporte (FPU13/01180). Carlos Gómez-Rodríguez is funded by an Oportunius program grant., Xunta de Galicia; R2014/034
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

LyS at TASS 2015: Deep Learning Experiments for Sentiment Analysis on Spanish Tweets

  • Vilares, David
  • Doval, Yerai
  • Alonso, Miguel A.
  • Gómez-Rodríguez, Carlos
[Abstract]: This paper describes the participation of the LyS group at tass 2015. In this year’s edition, we used a long short-term memory neural network to address the two proposed challenges: (1) sentiment analysis at a global level and (2) aspect-based sentiment analysis on football and political tweets. The performance of this deep learning approach is compared to our last-year model, based on a square-regularized logistic regression. Experimental results show that strategies such as unsupervised pre-training, sentiment-specific word embedding or modifying the current architec- ture might be needed to achieve state-of-the-art results., [Resumen]: Este art´ıculo describe la participaci´on del grupo LyS en el tass 2015. En la edici´on de este a˜no, hemos utilizado una red neuronal denominada long short- term memory para abordar los dos retos propuestos: (1) an´alisis del sentimiento a nivel global y (2) an´alisis del sentimiento a nivel de aspectos sobre tuits futbol´ısticos y de pol´ıtica. El rendimiento obtenido por esta red de aprendizaje profundo es comparado con el de nuestro sistema del a˜no pasado, una regresi´on log´ıstica con una regularizaci´on cuadr´atica. Los resultados experimentales muestran que es necesario incluir estrategias como pre-entrenamiento no supervisado, t´ecnicas espec´ıficas para representar palabras como vectores o modificar la arquitectura actual para alcanzar resultados acordes con el estado del arte., This research is supported by the Ministerio de Economía y Competitividad y FEDER (FFI2014-51978-C2) and Xunta de Galicia (R2014/034). The first author is funded by the Ministerio de Educación, Cultura y Deporte (FPU13/01180)., Xunta de Galicia; R2014/034
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

The megaphone of the people? Spanish SentiStrength for real-time analysis of political tweets

  • Vilares, David
  • Thelwall, Mike
  • Alonso, Miguel A.
This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/. This version of the article: Vilares, D., Thelwall, M., & Alonso, M. A. (2015). ‘The megaphone of the people? Spanish SentiStrength for real-time analysis of political tweets’ has been accepted for publication in Journal of Information Science, 41(6), 799-813. Copyright © 2015 The Authors. DOI: https://doi.org/10.1177/0165551515598926., [Abstract]: Twitter is an important platform for sharing opinions about politicians, parties and political decisions. These opinions can be exploited as a source of information to monitor the impact of politics on society. This article analyses the sentiment of 2,704,523 tweets referring to Spanish politicians and parties from a month in 2014–2015. The article makes three specific contributions: (a) enriching SentiStrength, a fast unsupervised sentiment strength detection system, for Spanish political tweeting; (b) analysing how linguistic phenomena such as negation, idioms and character duplication influence Spanish sentiment strength detection accuracy; and (c) analysing Spanish political tweets to rank political leaders, parties and personalities for popularity. Sentiment in Twitter for key politicians broadly reflects the main official polls for popularity but not for voting intention. In addition, the data suggests that the primary role of Twitter in politics is to select and amplify political events published by traditional media., This research is partially supported by the Ministerio de Economía y Competitividad (FFI2014-51978-C2) and Xunta de Galicia (R2014/034). The first author is partially funded by the Ministerio de Educación, Cultura y Deporte (FPU13/01180) and Inditex-UDC 2014 grants for international visits. The authors thank the annotators for help with this research., Xunta de Galicia; R2014/034
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

On the performance of phonetic algorithms in microtext normalization

  • Doval, Yerai
  • Vilares Ferro, Manuel
  • Vilares, Jesús
© 2018. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/. This version of the article: Doval, Y., Vilares, M. and Vilares, J. (2018) ‘On the performance of phonetic algorithms in microtext normalization’ has been accepted for publication in: Expert Systems with Applications, 113, pp. 213–222. The Version of Record is available online at: https://doi.org/10.1016/j.eswa.2018.07.016, [Abstract]: User–generated content published on microblogging social networks constitutes a priceless source of information. However, microtexts usually deviate from the standard lexical and grammatical rules of the language, thus making its processing by traditional intelligent systems very difficult. As an answer, microtext normalization consists in transforming those non–standard microtexts into standard well–written texts as a preprocessing step, allowing traditional approaches to continue with their usual processing. Given the importance of phonetic phenomena in non–standard text formation, an essential element of the knowledge base of a normalizer would be the phonetic rules that encode these phenomena, which can be found in the so–called phonetic algorithms. In this work we experiment with a wide range of phonetic algorithms for the English language. The aim of this study is to determine the best phonetic algorithms within the context of candidate generation for microtext normalization. In other words, we intend to find those algorithms that taking as input non–standard terms to be normalized allow us to obtain as output the smallest possible sets of normalization candidates which still contain the corresponding target standard words. As it will be stated, the choice of the phonetic algorithm will depend heavily on the capabilities of the candidate selection mechanism which we usually find at the end of a microtext normalization pipeline. The faster it can make the right choices among big enough sets of candidates, the more we can sacrifice on the precision of the phonetic algorithms in favour of coverage in order to increase the overall performance of the normalization system., This research has been partially funded by the Spanish Ministry of Economy, Industry and Competitiveness (MINECO) through projects TIN2017-85160-C2-1-R, TIN2017-85160-C2-2-R, FFI2014-51978-C2-1-R and FFI2014-51978-C2-2-R, and by the Autonomous Government of Galicia through projects ED431D-2017/12, ED431B-2017/01 and ED431D R2016/046. Moreover, Yerai Doval is funded by the Spanish State Secretariat for Research, Development and Innovation (which belongs to MINECO) and by the European Social Fund (ESF) under a FPI fellowship (BES-2015-073768) associated to project FFI2014-51978-C2-1-R., Xunta de Galicia; ED431D-2017/12, Xunta de Galicia; ED431B-2017/01, Xunta de Galicia; ED431D R2016/046
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Towards Robust Word Embeddings for Noisy Texts

  • Doval, Yerai
  • Vilares, Jesús
  • Gómez-Rodríguez, Carlos
[Abstract] Research on word embeddings has mainly focused on improving their performance on standard corpora, disregarding the difficulties posed by noisy texts in the form of tweets and other types of non-standard writing from social media. In this work, we propose a simple extension to the skipgram model in which we introduce the concept of bridge-words, which are artificial words added to the model to strengthen the similarity between standard words and their noisy variants. Our new embeddings outperform baseline models on noisy texts on a wide range of evaluation tasks, both intrinsic and extrinsic, while retaining a good performance on standard texts. To the best of our knowledge, this is the first explicit approach at dealing with these types of noisy texts at the word embedding level that goes beyond the support for out-of-vocabulary words., Yerai Doval has been supported by the Spanish Ministry of Economy, Industry and Competitiveness (MINECO) through the ANSWER-ASAP project (TIN2017-85160-C2-2-R); by the Spanish State Secretariat for Research, Development and Innovation (which belongs to MINECO) and the European Social Fund (ESF) through a FPI fellowship (BES-2015-073768) associated with TELEPARES project (FFI2014-51978-C2-1-R); and by the Xunta de Galicia through TELGALICIA research network (ED431D 2017/12). The work of Jesús Vilares and Carlos Gómez-Rodríguez has also been funded by MINECO through the ANSWER-ASAP project (TIN2017-85160-C2-1-R in this case); and by Xunta de Galicia through a Group with Potential for Growth grant (ED431B 2017/01), a Competitive Reference Group grant (ED431C 2020/11), and a Remarkable Research Centre grant for the CITIC research centre (ED431G/01), the latter co-funded by EU with ERDF funding. Finally, Carlos Gómez-Rodríguez has also received funding from the European Research Council (ERC), under the European Union’s Horizon 2020 research and innovation programme (FASTPARSE, Grant No. 714150), Xunta de Galicia; ED431D 2017/12, Xunta de Galicia; ED431B 2017/01, Xunta de Galicia; ED431C 2020/11, Xunta de Galicia; ED431G/01
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

A non-projective greedy dependency parser with bidirectional LSTMs

  • Vilares, David
  • Gómez-Rodríguez, Carlos
[Abstract]: The LyS-FASTPARSE team present BIST-COVINGTON, a neural implementation of the Covington (2001) algorithm for non-projective dependency parsing. The bidirectional LSTM approach by Kiperwasser and Goldberg (2016) is used to train a greedy parser with a dynamic oracle to mitigate error propagation. The model participated in the CoNLL 2017 UD Shared Task. In spite of not using any ensemble methods and using the baseline segmentation and PoS tagging, the parser obtained good results on both macro-average LAS and UAS in the big treebanks category (55 languages), ranking 7th out of 33 teams. In the all treebanks category (LAS and UAS) we ranked 16th and 12th. The gap between the all and big categories is mainly due to the poor performance on four parallel PUD treebanks, suggesting that some ‘suffixed’ treebanks (e.g. Spanish-AnCora) perform poorly on cross-treebank settings, which does not occur with the corresponding ‘unsuffixed’ treebank (e.g. Spanish). By changing that, we obtain the 11th best LAS among all runs (official and unofficial). The code is made available at https://github.com/CoNLL-UD-2017/LyS-FASTPARSE., David Vilares is funded by an FPU Grant 13/01180. Carlos Gómez-Rodríguez has received funding from the European Research Council (ERC), under the European Union’s Horizon 2020 research and innovation programme (FASTPARSE, grant agreement No 714150). Both authors have received funding from the TELEPARES-UDC project from MINECO.
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Comparing neural- and N-gram-based language models for word segmentation

  • Doval, Yerai
  • Gómez-Rodríguez, Carlos
[Abstract]: Word segmentation is the task of inserting or deleting word boundary characters in order to separate character sequences that correspond to words in some language. In this article we propose an approach based on a beam search algorithm and a language model working at the byte/character level, the latter component implemented either as an n-gram model or a recurrent neural network. The resulting system analyzes the text input with no word boundaries one token at a time, which can be a character or a byte, and uses the information gathered by the language model to determine if a boundary must be placed in the current position or not. Our aim is to use this system in a preprocessing step for a microtext normalization system. This means that it needs to effectively cope with the data sparsity present on this kind of texts. We also strove to surpass the performance of two readily available word segmentation systems: The well-known and accessible Word Breaker by Microsoft, and the Python module WordSegment by Grant Jenks. The results show that we have met our objectives, and we hope to continue to improve both the precision and the efficiency of our system in the future. © 2018 The Authors. Journal of the Association for Information Science and Technology published by Wiley Periodicals, Inc. on behalf of ASIS&T., This research received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation program (grant agreement no. 714150-FASTPARSE). It was partially funded by the Spanish Ministry of Economy and Competitiveness (MINECO) through projects FFI2014-51978-C2-1-R and FFI2014-51978-C2-2-R, and by the Autonomous Government of Galicia through both the Galician Network for Lexicography-RELEX (ED431D R2016/046) and Grant ED431B-2017/01. Moreover, Yerai Doval is funded by the Spanish State Secretariat for Research, Development and Innovation (which belongs to MINECO) and by the European Social Fund (ESF) under an FPI fellowship (BES-2015-073768) associated with project FFI2014-51978-C2-1-R. We gratefully acknowledge NVIDIA Corporation for the donation of a GTX Titan X GPU used for this research., Xunta de Galicia; ED431D R2016/046, Xunta e Galicia; ED431B-2017/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Universal, unsupervised (rule-based), uncovered sentiment analysis

  • Vilares, David
  • Gómez-Rodríguez, Carlos
  • Alonso, Miguel A.
© 2017. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/. This version of the article: Vilares, D., Gómez-Rodríguez, C. and Alonso, M.A. (2017) ‘Universal, unsupervised (rule-based), uncovered sentiment analysis’ has been accepted for publication in Knowledge-Based Systems, 118, pp. 45–55. The Version of Record is available online at https://doi.org/10.1016/j.knosys.2016.11.014., [Abstract]: We present a novel unsupervised approach for multilingual sentiment analysis driven by compositional syntax-based rules. On the one hand, we exploit some of the main advantages of unsupervised algorithms: (1) the interpretability of their output, in contrast with most supervised models, which behave as a black box and (2) their robustness across different corpora and domains. On the other hand, by introducing the concept of compositional operations and exploiting syntactic information in the form of universal dependencies, we tackle one of their main drawbacks: their rigidity on data that are structured differently depending on the language concerned. Experiments show an improvement both over existing unsupervised methods, and over state-of-the-art supervised models when evaluating outside their corpus of origin. Experiments also show how the same compositional operations can be shared across languages. The system is available at http://www.grupolys.org/software/UUUSA/, This research is supported by the Ministerio de Economía y Competitividad (FFI2014-51978-C2). David Vilares is funded by the Ministerio de Educación, Cultura y Deporte (FPU13/01180). Carlos Gómez-Rodríguez is funded by an Oportunius program grant (Xunta de Galicia). We thank Roman Klinger for his help in obtaining the German data.
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Sentiment Analysis on Monolingual, Multilingual and Code-Switching Twitter Corpora

  • Vilares, David
  • Alonso, Miguel A.
  • Gómez-Rodríguez, Carlos
[Abstract]: We address the problem of performing po- larity classification on Twitter over differ- ent languages, focusing on English and Spanish, comparing three techniques: (1) a monolingual model which knows the language in which the opinion is written, (2) a monolingual model that acts based on the decision provided by a language iden- tification tool and (3) a multilingual model trained on a multilingual dataset that does not need any language recognition step. Results show that multilingual models are even able to outperform the monolingual models on some monolingual sets. We introduce the first code-switching corpus with sentiment labels, showing the robust- ness of a multilingual approach., This research is supported by the Ministerio de Economía y Competitividad (FFI2014-51978-C2) and Xunta de Galicia (R2014/034). The first authoris funded by the Ministerio de Educación, Cultura y Deporte (FPU13/01180)., Xunta de Galicia; R2014/034
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

LyS at SemEval-2016 Task 4: Exploiting Neural Activation Values for Twitter Sentiment Classification and Quantification

  • Vilares, David
  • Doval, Yerai
  • Alonso, Miguel A.
  • Gómez-Rodríguez, Carlos
[Abstract]: In this paper we describe our deep learning approach for solving both two-, three- and fiveclass tweet polarity classification, and twoand five-class quantification. We first trained a convolutional neural network using pretrained Twitter word embeddings, so that we could extract the hidden activation values from the hidden layers once some input had been fed to the network. These values were then used as features for a support vector machine in both the classification and quantification subtasks, together with additional linguistic information in the former scenario. The results obtained for the classification subtasks show that this approach performs better than a single convolutional network, and for the quantification part it also yields good results. Official rankings locate us: 2nd (practically tied with 1st) for the binary classification task, 2nd for binary quantification and 4th (practically tied with 3rd) for the five-class polarity classification challenge., This research is supported by the Ministerio de Economía y Competitividad (FFI2014-51978-C2). David Vilares is funded by the Ministerio de Educación, Cultura y Deporte (FPU13/01180). Yerai Doval is funded by the Ministerio de Economía y Competitividad (BES-2015-073768). Carlos Gómez-Rodríguez is funded by an Oportunius program grant (Xunta de Galicia)
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Supervised sentiment analysis in multilingual environments

  • Vilares, David
  • Alonso, Miguel A.
  • Gómez-Rodríguez, Carlos
© 2017. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/. This version of the article: Vilares, D., Alonso, M.A. and Gómez-Rodríguez, C. (2017) ‘Supervised sentiment analysis in multilingual environments’ has been accepted for publication in Information Processing & Management, 53(3), pp. 595–607. The Version of Record is available online at https://doi.org/10.1016/j.ipm.2017.01.004., [Abstract]: This article tackles the problem of performing multilingual polarity classification on Twitter, comparing three techniques: (1) a multilingual model trained on a multilingual dataset, obtained by fusing existing monolingual resources, that does not need any language recognition step, (2) a dual monolingual model with perfect language detection on monolingual texts and (3) a monolingual model that acts based on the decision provided by a language identification tool. The techniques were evaluated on monolingual, synthetic multilingual and code-switching corpora of English and Spanish tweets. In the latter case we introduce the first code-switching Twitter corpus with sentiment labels. The samples are labelled according to two well-known criteria used for this purpose: the SentiStrength scale and a trinary scale (positive, neutral and negative categories). The experimental results show the robustness of the multilingual approach (1) and also that it outperforms the monolingual models on some monolingual datasets., This research was supported by the Ministerio de Economía y Competitividad (FFI2014-51978-C2) and Xunta de Galicia (R2014/034). David Vilares is funded by the Ministerio de Educación, Cultura y Deporte (FPU13/01180). Carlos Gómez-Rodríguez is funded by an Oportunius program grant (Xunta de Galicia)., Xunta de Galicia; R2014/034