Pasar al contenido principal

AVANCES EN NUEVOS SISTEMAS DE EXTRACCION DE RESPUESTAS CON ANALISIS SEMANTICO Y APRENDIZAJE PROFUNDO

TIN2017-85160-C2-2-R

•
Nombre agencia financiadora Agencia Estatal de Investigación
Acrónimo agencia financiadora AEI
Programa Programa Estatal de I+D+i Orientado a los Retos de la Sociedad
Convocatoria Proyectos I+D+i Retos de Investigación
Año convocatoria 2017
Plan Estatal 2017-2020 Plan Estatal de Investigación Científica y Técnica y de Innovación
Centro beneficiario UNIVERSIDAD DE VIGO

Resultados relacionados

Resultados totales (Incluyendo duplicados): 9
Encontrada(s) 1 página(s)

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 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)

Adaptive scheduling for adaptive sampling in pos taggers construction

  • Vilares Ferro, Manuel
  • Darriba Bilbao, Víctor Manuel
  • Vilares Ferro, Jesús
We introduce an adaptive scheduling for adaptive sampling as a novel way of machine learning in the construction of part-of-speech taggers. The goal is to speed up the training on large data sets, without significant loss of performance with regard to an optimal configuration. In contrast to previous methods using a random, fixed or regularly rising spacing between the instances, ours analyzes the shape of the learning curve geometrically in conjunction with a functional model to increase or decrease it at any time. The algorithm proves to be formally correct regarding our working hypotheses. Namely, given a case, the following one is the nearest ensuring a net gain of learning ability from the former, it being possible to modulate the level of requirement for this condition. We also improve the robustness of sampling by paying greater attention to those regions of the training data base subject to a temporary inflation in performance, thus preventing the learning from stopping prematurely. The proposal has been evaluated on the basis of its reliability to identify the convergence of models, corroborating our expectations. While a concrete halting condition is used for testing, users can choose any condition whatsoever to suit their own specific needs., Agencia Estatal de Investigación | Ref. TIN2017-85160-C2-1-R, Agencia Estatal de Investigación | Ref. TIN2017-85160-C2-2-R, Xunta de Galicia | Ref. ED431C 2018/50, Xunta de Galicia | Ref. ED431D 2017/12
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Absolute convergence and error thresholds in non-active adaptive sampling

  • Vilares Ferro, Manuel
  • Darriba Bilbao, Víctor Manuel
  • Vilares Ferro, Jesús
Financiado para publicación en acceso aberto: Universidade de Vigo/CISUG, Non-active adaptive sampling is a way of building machine learning models from a training data base which are supposed to dynamically and automatically derive guaranteed sample size. In this context and regardless of the strategy used in both scheduling and generating of weak predictors, a proposal for calculating absolute convergence and error thresholds is described. We not only make it possible to establish when the quality of the model no longer increases, but also supplies a proximity condition to estimate in absolute terms how close it is to achieving such a goal, thus supporting decision making for fine-tuning learning parameters in model selection. The technique proves its correctness and completeness with respect to our working hypotheses, in addition to strengthening the robustness of the sampling scheme. Tests meet our expectations and illustrate the proposal in the domain of natural language processing, taking the generation of part-of-speech taggers as case study., Agencia Estatal de Investigación | Ref. TIN2017-85160-C2-1-R, Agencia Estatal de Investigación | Ref. TIN2017-85160-C2-2-R, Agencia Estatal de Investigación | PID2020-113230RB-C21, Agencia Estatal de Investigación | PID2020-113230RB-C22, Xunta de Galicia | Ref. ED431C 2018/50
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Capítulo o parte de un libro (BookPart)

Parallel corpora as a resource in the translation process: an example within the biomedical field

  • Sánchez Trigo, Maria Elena
  • Varela Vila, Tamara
This paper presents the design, processing and potential exploitation of a parallel biomedical corpus in French, English and Spanish about orphan drugs, ORPHACOR. This open access corpus, which incorporates two text genres (summaries of product characteristics and package leaflets), is intended for translators, interpreters and related professionals. It fulfills a dual purpose: 1) to compile relevant texts on the treatment of rare diseases, a multidisciplinary subdomain of medicine in which there is social demand for information dissemination, and 2) to provide data on the characteristic terminological usages within the selected text genres. The corpus has been integrated into RERCOR, the portal of multilingual language resources on rare diseases, in development since 2018. The features of the portal oÁer the possibility of making a range of queries through the application with diÁerent parameters to obtain useful data adapted to the users’ individual needs, Xunta de Galicia | Ref. ED431C 2018/50, Agencia Estatal de Investigación | Ref. TIN2017-85160-C2-2-R
RUA. Repositorio Institucional de la Universidad de Alicante
Publicación Otros (Other)

Procesamiento de Expresiones Multipalabra en gallego mediante Aprendizaje Profundo, Multiword expressions processing in Galician using Deep Learning

  • Darriba Bilbao, Víctor Manuel
  • Doval, Yerai
  • Kuriyozov, Elmurod
El tratamiento de Expresiones Multipalabra es todavía una tarea pendiente en el Procesamiento del Lenguaje Natural. En este trabajo pretendemos determinar experimentalmente la utilidad de los modelos de Aprendizaje Automático para el procesamiento de Expresiones Multipalabra en gallego. Para ello usamos CORGA, un corpus con 40 millones de palabras, con el cual entrenamos modelos transformer de Aprendizaje Profundo, y comparamos su rendimiento con el de modelos más tradicionales de campo aleatorio condicional.
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