Pasar al contenido principal

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

TIN2017-85160-C2-1-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 UNIVERSIDADE DA CORUÑA

Resultados relacionados

Resultados totales (Incluyendo duplicados): 40
Encontrada(s) 1 página(s)

Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Faster shift-reduce constituent parsing with a non-binary, bottom-up strategy

  • Fernández González, Daniel
  • Gómez Rodríguez, Carlos
An increasingly wide range of artificial intelligence applications rely on syntactic information to process and extract meaning from natural language text or speech, with constituent trees being one of the most widely used syntactic formalisms. To produce these phrase-structure representations from sentences in natural language, shift-reduce constituent parsers have become one of the most efficient approaches. Increasing their accuracy and speed is still one of the main objectives pursued by the research community so that artificial intelligence applications that make use of parsing outputs, such as machine translation or voice assistant services, can improve their performance. With this goal in mind, we propose in this article a novel non-binary shift-reduce algorithm for constituent parsing. Our parser follows a classical bottom-up strategy but, unlike others, it straightforwardly creates non-binary branchings with just one transition, instead of requiring prior binarization or a sequence of binary transitions, allowing its direct application to any language without the need of further resources such as percolation tables. As a result, it uses fewer transitions per sentence than existing transition-based constituent parsers, becoming the fastest such system and, as a consequence, speeding up downstream applications. Using static oracle training and greedy search, the accuracy of this novel approach is on par with state-of-the-art transition-based constituent parsers and outperforms all top-down and bottom-up greedy shift-reduce systems on the Wall Street Journal section from the English Penn Treebank and the Penn Chinese Treebank. Additionally, we develop a dynamic oracle for training the proposed transition-based algorithm, achieving further improvements in both benchmarks and obtaining the best accuracy to date on the Penn Chinese Treebank among greedy shift-reduce parsers., Agencia Estatal de Investigación | Ref. TIN2017-85160-C2-1-R, Xunta de Galicia | Ref. ED431B 2017/01
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 Contribución de congreso (ConferenceOutput)

Supervised classifiers to identify hate speech on English and Spanish tweets

  • Almatarneh, Sattam
  • Gamallo, Pablo
  • Ribadas Pena, Francisco José
  • Alexeev, Alexey
Consistently with social and political concern about hatred and harassment through social media, in recent years, automatic hate-speech detection and offensive behavior in social media are gaining a lot of attention. In this paper, we examine the performance of several supervised classifiers in the process of identifying hate speech on Twitter. More precisely, we do an empirical study that analyzes the influence of two types of linguistic features (n-grams, word embeddings) when they are used to feed different supervised machine learning classifiers: Support Vector Machine (SVM), Gaussian Naive Bayes (GNB), Complement Naive Bayes (CNB), Decision Tree (DT), Nearest Neighbors (KN), Random Forest (RF) and Neural Network (NN). The experiments we have carried out show that CNB, SVM, and RF are better than the rest classifiers in English and Spanish languages by taking into account all features., Xunta de Galicia | Ref. ED431C 2018/50, Agencia Estatal de Investigación | Ref. TIN2017-85160-C2-2-R
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Multitask pointer network for multi-representational parsing

  • Fernández González, Daniel
  • Gómez Rodríguez, Carlos
Dependency and constituent trees are widely used by many artificial intelligence applications for representing the syntactic structure of human languages. Typically, these structures are separately produced by either dependency or constituent parsers. In this article, we propose a transition-based approach that, by training a single model, can efficiently parse any input sentence with both constituent and dependency trees, supporting both continuous/projective and discontinuous/non-projective syntactic structures. To that end, we develop a Pointer Network architecture with two separate task-specific decoders and a common encoder, and follow a multitask learning strategy to jointly train them. The resulting quadratic system, not only becomes the first parser that can jointly produce both unrestricted constituent and dependency trees from a single model, but also proves that both syntactic formalisms can benefit from each other during training, achieving state-of-the-art accuracies in several widely-used benchmarks such as the continuous English and Chinese Penn Treebanks, as well as the discontinuous German NEGRA and TIGER datasets., Agencia Estatal de Investigación | Ref. TIN2017-85160-C2-1-R, Agencia Estatal de Investigación | Ref. PID2020-113230RB-C21, Xunta de Galicia | Ref. ED431C 2020/11, Xunta de Galicia | Ref. ED431G 2019/01, Universidade da Coruña/CISUG
RUA. Repositorio Institucional de la Universidad de Alicante
Publicación Otros (Other)

Discovering topics in Twitter about the COVID-19 outbreak in Spain, Descubriendo temas en Twitter sobre el brote del COVID-19 en España

  • Agüero-Torales, Marvin M.
  • Vilares Calvo, David
  • López-Herrera, Antonio G.
In this work, we apply topic modeling to study what users have been discussing in Twitter during the beginning of the COVID-19 pandemic. More particularly, we explore the period of time that includes three differentiated phases of the COVID-19 crisis in Spain: the pre-crisis time, the outbreak, and the beginning of the lockdown. To do so, we first collect a large corpus of Spanish tweets and clean them. Then, we cluster the tweets into topics using a Latent Dirichlet Allocation model, and define generative and discriminative routes to later extract the most relevant keywords and sentences for each topic. Finally, we provide an exhaustive qualitative analysis about how such topics correspond to the situation in Spain at different stages of the crisis.
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Not All Linearizations Are Equally Data-Hungry in Sequence Labeling Parsing

  • Muñoz-Ortiz, Alberto
  • Strzyz, Michalina
  • Vilares, David
It was held online, 1-3 September 2021., [Absctract]: Different linearizations have been proposed to cast dependency parsing as sequence labeling and solve the task as: (i) a head selection problem, (ii) finding a representation of the token arcs as bracket strings, or (iii) associating partial transition sequences of a transition-based parser to words. Yet, there is little understanding about how these linearizations behave in low-resource setups. Here, we first study their data efficiency, simulating data-restricted setups from a diverse set of rich-resource treebanks. Second, we test whether such differences manifest in truly low-resource setups. The results show that head selection encodings are more data-efficient and perform better in an ideal (gold) framework, but that such advantage greatly vanishes in favour of bracketing formats when the running setup resembles a real-world low-resource configuration., This work is supported by a 2020 Leonardo Grant for Researchers and Cultural Creators from the FBBVA. The work also receives funding from the European Research Council (FASTPARSE, grant agreement No 714150), from ERDF/MICINNAEI (ANSWER-ASAP, TIN2017-85160-C2-1- R, SCANNER, PID2020-113230RB-C21), from Xunta de Galicia (ED431C 2020/11), and from Centro de Investigacion de Galicia ‘CITIC’, funded ´ by Xunta de Galicia and the European Union (European Regional Development Fund- Galicia 2014- 2020 Program) by grant ED431G 2019/01., Xunta de Galicia; ED431C 2020/11, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Grounding the Semantics of Part-of-Day Nouns Worldwide using Twitter

  • Vilares, David
  • Gómez-Rodríguez, Carlos
Held 6 June 2018, New Orleans, Louisiana, USA., [Absctract]: The usage of part-of-day nouns, such as ‘night’, and their time-specific greetings (‘good night’), varies across languages and cultures. We show the possibilities that Twitter offers for studying the semantics of these terms and its variability between countries. We mine a worldwide sample of multilingual tweets with temporal greetings, and study how their frequencies vary in relation with local time. The results provide insights into the semantics of these temporal expressions and the cultural and sociological factors influencing their usage., DV and CGR receive funding from the European Research Council (ERC), under the European Union’s Horizon 2020 research and innovation programme (FASTPARSE, grant agreement No 714150), from the TELEPARESUDC project (FFI2014-51978-C2-2-R) and the ANSWER-ASAP project (TIN2017-85160-C2-1- R) from MINECO, and from Xunta de Galicia (ED431B 2017/01)., Xunta de Galicia; ED431B 2017/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

On the Use of Parsing for Named Entity Recognition

  • Alonso, Miguel A.
  • Gómez-Rodríguez, Carlos
  • Vilares, Jesús
[Abstract] Parsing is a core natural language processing technique that can be used to obtain the structure underlying sentences in human languages. Named entity recognition (NER) is the task of identifying the entities that appear in a text. NER is a challenging natural language processing task that is essential to extract knowledge from texts in multiple domains, ranging from financial to medical. It is intuitive that the structure of a text can be helpful to determine whether or not a certain portion of it is an entity and if so, to establish its concrete limits. However, parsing has been a relatively little-used technique in NER systems, since most of them have chosen to consider shallow approaches to deal with text. In this work, we study the characteristics of NER, a task that is far from being solved despite its long history; we analyze the latest advances in parsing that make its use advisable in NER settings; we review the different approaches to NER that make use of syntactic information; and we propose a new way of using parsing in NER based on casting parsing itself as a sequence labeling task., Xunta de Galicia; ED431C 2020/11, Xunta de Galicia; ED431G 2019/01, This work has been funded by MINECO, AEI and FEDER of UE through the ANSWER-ASAP project (TIN2017-85160-C2-1-R); and by Xunta de Galicia through a Competitive Reference Group grant (ED431C 2020/11). CITIC, as Research Center of the Galician University System, is funded by the Consellería de Educación, Universidade e Formación Profesional of the Xunta de Galicia through the European Regional Development Fund (ERDF/FEDER) with 80%, the Galicia ERDF 2014-20 Operational Programme, and the remaining 20% from the Secretaría Xeral de Universidades (Ref. ED431G 2019/01). 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).
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Sentiment Analysis for Fake News Detection

  • Alonso, Miguel A.
  • Vilares, David
  • Gómez-Rodríguez, Carlos
  • Vilares, Jesús
[Abstract] In recent years, we have witnessed a rise in fake news, i.e., provably false pieces of information created with the intention of deception. The dissemination of this type of news poses a serious threat to cohesion and social well-being, since it fosters political polarization and the distrust of people with respect to their leaders. The huge amount of news that is disseminated through social media makes manual verification unfeasible, which has promoted the design and implementation of automatic systems for fake news detection. The creators of fake news use various stylistic tricks to promote the success of their creations, with one of them being to excite the sentiments of the recipients. This has led to sentiment analysis, the part of text analytics in charge of determining the polarity and strength of sentiments expressed in a text, to be used in fake news detection approaches, either as a basis of the system or as a complementary element. In this article, we study the different uses of sentiment analysis in the detection of fake news, with a discussion of the most relevant elements and shortcomings, and the requirements that should be met in the near future, such as multilingualism, explainability, mitigation of biases, or treatment of multimedia elements., Xunta de Galicia; ED431G 2019/01, Xunta de Galicia; ED431C 2020/11, This work has been funded by FEDER/Ministerio de Ciencia, Innovación y Universidades — Agencia Estatal de Investigación through the ANSWERASAP project (TIN2017-85160-C2-1-R); and by Xunta de Galicia through a Competitive Reference Group grant (ED431C 2020/11). CITIC, as Research Center of the Galician University System, is funded by the Consellería de Educación, Universidade e Formación Profesional of the Xunta de Galicia through the European Regional Development Fund (ERDF/FEDER) with 80%, the Galicia ERDF 2014-20 Operational Programme, and the remaining 20% from the Secretaría Xeral de Universidades (ref. ED431G 2019/01). David Vilares is also supported by a 2020 Leonardo Grant for Researchers and Cultural Creators from the BBVA Foundation. 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)
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Isolation Number versus Domination Number of Trees

  • Lemańska, Magdalena
  • Souto Salorio, María José
  • Dapena, Adriana
  • Vázquez Araújo, Francisco Javier
[Abstract] If 𝐺 = (Vɢ,Eɢ) is a graph of order n, we call 𝑆 ⊆ Vɢ an isolating set if the graph induced by Vɢ − Nɢ[𝑆] contains no edges. The minimum cardinality of an isolating set of 𝐺 is called the isolation number of 𝐺, and it is denoted by 𝜄(𝐺). It is known that 𝜄(𝐺) ≤ ⁿ⁄₃ and the bound is sharp. A subset 𝑆 ⊆ Vɢ is called dominating in 𝐺 if Nɢ[𝑆] = Vɢ. The minimum cardinality of a dominating set of 𝐺 is the domination number, and it is denoted by 𝛾(𝐺). In this paper, we analyze a family of trees 𝑇 where 𝜄(𝑇) = 𝛾(𝑇), and we prove that 𝜄(T) = ⁿ⁄₃ implies 𝜄(𝑇) = 𝛾(𝑇). Moreover, we give different equivalent characterizations of such graphs and we propose simple algorithms to build these trees from the connections of stars., CITIC, as Research Center accredited by Galician University System, is funded by "Consellería de Cultura, Educación e Universidade from Xunta de Galicia", supported in an 80% through ERDF Funds, ERDF Operational Programme Galicia 2014-2020, and the remaining 20% by "Secretaría Xeral de Universidades (Grant ED431G 2019/01). This research was also funded by Agencia Estatal de Investigación of Spain (PID2019-104958RB-C42 and TIN2017-85160-C2-1-R) and ERDF funds of the EU (AEI/FEDER, UE)., Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Bracketing Encodings for 2-Planar Dependency Parsing

  • Strzyz, Michalina
  • Vilares, David
  • Gómez-Rodríguez, Carlos
Held online due to COVID-19. December 2020, Barcelona, Spain, [Absctract]: We present a bracketing-based encoding that can be used to represent any 2-planar dependency tree over a sentence of length n as a sequence of n labels, hence providing almost total coverage of crossing arcs in sequence labeling parsing. First, we show that existing bracketing encodings for parsing as labeling can only handle a very mild extension of projective trees. Second, we overcome this limitation by taking into account the well-known property of 2-planarity, which is present in the vast majority of dependency syntactic structures in treebanks, i.e., the arcs of a dependency tree can be split into two planes such that arcs in a given plane do not cross. We take advantage of this property to design a method that balances the brackets and that encodes the arcs belonging to each of those planes, allowing for almost unrestricted non-projectivity (∼99.9% coverage) in sequence labeling parsing. The experiments show that our linearizations improve over the accuracy of the original bracketing encoding in highly non-projective treebanks (on average by 0.4 LAS), while achieving a similar speed. Also, they are especially suitable when PoS tags are not used as input parameters to the models., This work has received funding from the European Research Council (ERC), which has funded this research under the European Union’s Horizon 2020 research and innovation programme (FASTPARSE, grant agreement No 714150), from MINECO (ANSWER-ASAP, TIN2017-85160-C2-1-R), from Xunta de Galicia (ED431C 2020/11), and from Centro de Investigacion de Galicia ‘CITIC’, funded by Xunta de ´ Galicia and the European Union (European Regional Development Fund- Galicia 2014-2020 Program), by grant ED431G 2019/01. DV is supported by a 2020 Leonardo Grant for Researchers and Cultural Creators from the BBVA Foundation., Xunta de Galicia; ED431C 2020/11, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Global Transition-based Non-projective Dependency Parsing

  • Fernández-González, Daniel
  • Shi, Tianze
  • Lee, Lillian
56th Annual Meeting of the Association for Computational Linguistics, ACL 2018. Melbourne15 July - 20 July 2018., [Absctract]: Shi, Huang, and Lee (2017a) obtained state-of-the-art results for English and Chinese dependency parsing by combining dynamic-programming implementations of transition-based dependency parsers with a minimal set of bidirectional LSTM features. However, their results were limited to projective parsing. In this paper, we extend their approach to support non-projectivity by providing the first practical implementation of the MH4 algorithm, an Opn4q mildly non projective dynamic-programming parser with very high coverage on non-projective treebanks. To make MH4 compatible with minimal transition-based feature sets, we introduce a transition-based interpretation of it in which parser items are mapped to sequences of transitions. We thus obtain the first implementation of global decoding for non-projective transition-based parsing, and demonstrate empirically that it is more effective than its projective counterpart in parsing a number of highly non-projective languages., We thank the three anonymous reviewers for their helpful comments. CG 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), from the TELEPARESUDC project (FFI2014-51978-C2-2-R) and the ANSWER-ASAP project (TIN2017-85160-C2-1-R) from MINECO, and from Xunta de Galicia (ED431B 2017/01). TS and LL were supported in part by a Google Focused Research Grant to Cornell University. LL was also supported in part by NSF grant SES-1741441. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation or other sponsors ., Xunta de Galicia; ED431B 2017/01, Estados Unidos. National Science Foundation; SES-1741441
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Exploring cross-lingual word embeddings for the inference of bilingual dictionaries

  • García, Marcos
  • García Salido, Marcos
  • Alonso, Miguel A.
Paper included in: Proceedings of TIAD-2019 Shared Task – Translation Inference Across Dictionaries co-located with the 2nd Language, Data and Knowledge Conference (LDK 2019) Leipzig, Germany, May 20, 2019., [Abstract]: We describe four systems to generate automatically bilingual dictionaries based on existing ones: three transitive systems differing only in the pivot language used, and a system based on a different approach which only needs monolingual corpora in both the source and target languages. All four methods make use of cross-lingual word embeddings trained on monolingual corpora, and then mapped into a shared vector space. Experimental results confirm that our strategy has a good coverage and recall, achieving a performance comparable to to the best submitted systems on the TIAD 2019 gold standard set among the teams participating at the TIAD shared task., Research supported by a 2017 Leonardo Grant for Researchers and Cultural Creators (BBVA Foundation), by Ministerio de Economía, Industria y Competitividad (Grant TIN2017-85160-C2-1-R), and by Xunta de Galicia (Grant ED431B-2017/01). Marcos Garcia has been funded by a Juan de la Cierva grant (IJCI-2016-29598), and Marcos García-Salido by a post-doctoral grant from Xunta de Galicia (ED481D 2017/009). We gratefully acknowledge the support of NVIDIA Corporation with the donation of the Titan Xp GPU used for this research., Xunta de Galicia; ED481D 2017/009, Xunta de Galicia; ED431B-2017/01
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Harry Potter and the Action Prediction Challenge from Natural Language

  • Vilares, David
  • Gómez-Rodríguez, Carlos
The 17th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT 2019) was held in Minneapolis from June 2nd to June 7th, 2019., [Absctract]: We explore the challenge of action prediction from textual descriptions of scenes, a testbed to approximate whether text inference can be used to predict upcoming actions. As a case of study, we consider the world of the Harry Potter fantasy novels and inferring what spell will be cast next given a fragment of a story. Spells act as keywords that abstract actions (e.g. ‘Alohomora’ to open a door) and denote a response to the environment. This idea is used to automatically build HPAC, a corpus containing 82,836 samples and 85 actions. We then evaluate different baselines. Among the tested models, an LSTM-based approach obtains the best performance for frequent actions and large scene descriptions, but approaches such as logistic regression behave well on infrequent actions., This work has received support from the TELEPARES-UDC project (FFI2014-51978-C2- 2-R) and the ANSWER-ASAP project (TIN2017- 85160-C2-1-R) from MINECO, and from Xunta de Galicia (ED431B 2017/01), and from the European Research Council (ERC), under the European Union’s Horizon 2020 research and innovation programme (FASTPARSE, grant agreement No 714150), Xunta de Galicia; ED431B 2017/01
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

On the Logistical Difficulties and Findings of Jopara Sentiment Analysis

  • Agüero-Torales, Marvin M.
  • Vilares, David
  • López-Herrera, Antonio G.
[Abstract] This paper addresses the problem of sentiment analysis for Jopara, a code-switching language between Guarani and Spanish. We first collect a corpus of Guarani-dominant tweets and discuss on the difficulties of finding quality data for even relatively easy-to-annotate tasks, such as sentiment analysis. Then, we train a set of neural models, including pre-trained language models, and explore whether they perform better than traditional machine learning ones in this low-resource setup. Transformer architectures obtain the best results, despite not considering Guarani during pre-training, but traditional machine learning models perform close due to the low-resource nature of the problem., DV is supported by a 2020 Leonardo Grant for Researchers and Cultural Creators from the FBBVA. 15 DV also receives funding from MINECO (ANSWER-ASAP, TIN2017-85160-C2-1-R), from Xunta de Galicia (ED431C 2020/11), from Centro de Investigación de Galicia ‘CITIC’, funded by Xunta de Galicia and the European Union (European Regional Development Fund- Galicia 2014-2020 Program) by grant ED431G 2019/01, Xunta de Galicia; ED431C 2020/11, Xunta de Galicia; ED431G 2019/01, https://aclanthology.org/2021.calcs-1
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Bertinho: Galician BERT Representations, Bertinho: Representaciones BERT para el gallego

  • Vilares, David
  • García, Marcos
  • Gómez-Rodríguez, Carlos
[Abstract]: This paper presents a monolingual BERT model for Galician. We follow the recent trend that shows that it is feasible to build robust monolingual BERT models even for relatively low-resource languages, while performing better than the well-known official multilingual BERT (mBERT). More particularly, we release two monolingual Galician BERT models, built using 6 and 12 transformer layers, respectively; trained with limited resources (∼45 million tokens on a single GPU of 24GB). We then provide an exhaustive evaluation on a number of tasks such as POS-tagging, dependency parsing and named entity recognition. For this purpose, all these tasks are cast in a pure sequence labeling setup in order to run BERT without the need to include any additional layers on top of it (we only use an output classification layer to map the contextualized representations into the predicted label). The experiments show that our models, especially the 12-layer one, outperform the results of mBERT in most tasks., [Resumen]: Este artículo presenta un modelo BERT monolingüe para el gallego. Nos basamos en la tendencia actual que ha demostrado que es posible crear modelos BERT monolingües robustos incluso para aquellos idiomas para los que hay una relativa escasez de recursos, funcionando éstos mejor que el modelo BERT multilingüe oficial (mBERT). Concretamente, liberamos dos modelos monolingües para el gallego, creados con 6 y 12 capas de transformers, respectivamente, y entrenados con una limitada cantidad de recursos (~45 millones de palabras sobre una única GPU de 24GB.) Para evaluarlos realizamos un conjunto exhaustivo de experimentos en tareas como análisis morfosintáctico, análisis sintáctico de dependencias o reconocimiento de entidades. Para ello, abordamos estas tareas como etiquetado de secuencias, con el objetivo de ejecutar los modelos BERT sin la necesidad de incluir ninguna capa adicional (únicamente se a~nade la capa de salida encargada de transformar las representaciones contextualizadas en la etiqueta predicha). Los experimentos muestran que nuestros modelos, especialmente el de 12 capas, mejoran los resultados de mBERT en la mayor parte de las tareas., This work has received funding from the European Research Council (ERC), which has funded this research under the European Union's Horizon 2020 research and innovation programme (FASTPARSE, grant agreement No 714150), from MINECO (ANSWER-ASAP, TIN2017-85160-C2-1-R), from Xunta de Galicia (ED431C 2020/11), from Centro de Investigación de Galicia `CITIC', funded by Xunta de Galicia and the European Union (European Regional Development Fund- Galicia 2014-2020 Program), by grant ED431G 2019/01, and by Centro Singular de Investigación en Tecnoloxías Intelixentes (CiTIUS), ERDF 2014-2020: Call ED431G 2019/04. DV is supported by a 2020 Leonardo Grant for Researchers and Cultural Creators from the BBVA Foundation. MG is supported by a Ramón y Cajal grant (RYC2019-028473-I)., Xunta de Galicia; ED431C 2020/11, Xunta de Galicia; ED431G 2019/01, Xunta de Galicia; D431G 2019/04
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Construction and evaluation of sentiment Datasets for low-resource languages: the case of Uzbek

  • Kuriyozov, Elmurod
  • Matlatipov, Sanatbek
  • Alonso, Miguel A.
  • Gómez-Rodríguez, Carlos
This is the Author Accepted Manuscript. This version of the conference paper has been accepted for publication, after peer review and is subject to Springer Nature’s AM terms of use (https://www.springernature.com/gp/open-science/policies/accepted-manuscript-terms), but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: https://doi.org/10.1007/978-3-031-05328-3_15., Conference paper presented at: 9th Language and Technology Conference, LTC 2019, Poznan, Poland, May 17–19, 2019., [Abstract]: To our knowledge, the majority of human language processing technologies for low-resource languages don’t have well-established linguistic resources for the development of sentiment analysis applications. Therefore, it is in dire need of such tools and resources to overcome the NLP barriers, so that, low-resource languages can deliver more benefits. In this paper, we fill that gap by providing its first annotated corpora for Uzbek language polarity classification. Our methodology considers collecting a medium-size manually annotated dataset and a larger-size dataset automatically translated from existing resources. Then, we use these datasets to train what, to our knowledge, are the first sentiment analysis models on the Uzbek language, using both traditional machine learning techniques and recent deep learning models. Both sets of techniques achieve similar accuracy (the best model on the manually annotated test set is a convolutional neural network with 88.89% accuracy, and on the translated set, a logistic regression with 89.56% accuracy); with the accuracy of the deep learning models being limited by the quality of available pre-trained word embeddings., This work has received funding from ERDF/MICINN-AEI (ANSWER-ASAP, TIN2017-85160-C2-1-R; SCANNER-UDC, PID2020-113230RB-C21), from Xunta de Galicia (ED431C 2020/11), and from Centro de Investigación de Galicia “CITIC”, funded by Xunta de Galicia and the European Union (ERDF - Galicia 2014–2020 Program), by grant ED431G 2019/01. Elmurod Kuriyozov was funded for his PhD by El-Yurt-Umidi Foundation under the Cabinet of Ministers of the Republic of Uzbekistan., Xunta de Galicia; ED431C 2020/11, Xunta de Galicia; ED431G 2019/01
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 Contribución de congreso (ConferenceOutput)

Left-to-Right Dependency Parsing with Pointer Networks

  • Fernández-González, Daniel
  • Gómez-Rodríguez, Carlos
Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), [Abstract]: We propose a novel transition-based algorithm that straightforwardly parses sentences from left to right by building n attachments, with n being the length of the input sentence. Similarly to the recent stack-pointer parser by Ma et al. (2018), we use the pointer network framework that, given a word, can directly point to a position from the sentence. However, our left-to-right approach is simpler than the original top-down stack-pointer parser (not requiring a stack) and reduces transition sequence length in half, from 2n-1 actions to n. This results in a quadratic non-projective parser that runs twice as fast as the original while achieving the best accuracy to date on the English PTB dataset (96.04% UAS, 94.43% LAS) among fully-supervised single-model dependency parsers, and improves over the former top-down transition system in the majority of languages tested., This work 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), from MINECO (FFI2014-51978-C2-2-R, TIN2017-85160-C2-1-R) and from Xunta de Galicia (ED431B 2017/01)., Xunta de Galicia; ED431B 2017/01
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Sequence Labeling Parsing by Learning across Representations

  • Strzyz, Michalina
  • Vilares, David
  • Gómez-Rodríguez, Carlos
The 57th Annual Meeting of the Association for Computational Linguistics (ACL) took place in Florence (Italy) at the 'Fortezza da Basso' from July 28th to August 2nd, 2019., [Absctract]: We use parsing as sequence labeling as a common framework to learn across constituency and dependency syntactic abstractions. To do so, we cast the problem as multitask learning (MTL). First, we show that adding a parsing paradigm as an auxiliary loss consistently improves the performance on the other paradigm. Secondly, we explore an MTL sequence labeling model that parses both representations, at almost no cost in terms of performance and speed. The results across the board show that on average MTL models with auxiliary losses for constituency parsing outperform single-task ones by 1.05 F1 points, and for dependency parsing by 0.62 UAS points., This work 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), from the ANSWER-ASAP project (TIN2017-85160-C2-1-R) from MINECO, and from Xunta de Galicia (ED431B 2017/01)., Xunta de Galicia; ED431B 2017/01
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Artificially Evolved Chunks for Morphosyntactic Analysis

  • Anderson, Mark
  • Vilares, David
  • Gómez-Rodríguez, Carlos
It was held on the last week of August (Aug 26-30), in the center of Paris, in the "Grand amphitheatre du Monde anglophone" of the Sorbonne Nouvelle., [Absctract]: We introduce a language-agnostic evolutionary technique for automatically extracting chunks from dependency treebanks. We evaluate these chunks on a number of morphosyntactic tasks, namely POS tagging, morphological feature tagging, and dependency parsing. We test the utility of these chunks in a host of different ways. We first learn chunking as one task in a shared multitask framework together with POS and morphological feature tagging. The predictions from this network are then used as input to augment sequence-labelling dependency parsing. Finally, we investigate the impact chunks have on dependency parsing in a multi-task framework. Our results from these analyses show that these chunks improve performance at different levels of syntactic abstraction on English UD treebanks and a small, diverse subset of non-English UD treebanks., This work has received funding from the European Research Council (ERC), under the European Unions Horizon 2020 research and innovation programme (FASTPARSE, grant agreement No 714150), from the ANSWER-ASAP project (TIN2017-85160-C2-1-R) from MINECO, and from Xunta de Galicia (ED431B 2017/01). We thank one anonymous reviewer for in-depth comments and suggestions., Xunta de Galicia; ED431B 2017/01
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Viable Dependency Parsing as Sequence Labeling

  • Strzyz, Michalina
  • Vilares, David
  • Gómez-Rodríguez, Carlos
[Abstract]: We recast dependency parsing as a sequence labeling problem, exploring several encodings of dependency trees as labels. While dependency parsing by means of sequence labeling had been attempted in existing work, results suggested that the technique was impractical. We show instead that with a conventional BILSTM-based model it is possible to obtain fast and accurate parsers. These parsers are conceptually simple, not needing traditional parsing algorithms or auxiliary structures. However, experiments on the PTB and a sample of UD treebanks show that they provide a good speed-accuracy tradeoff, with results competitive with more complex approaches., This work 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), from the TELEPARESUDC project (FFI2014-51978-C2-2-R) and the ANSWER-ASAP project (TIN2017-85160-C2-1-R) from MINECO, and from Xunta de Galicia (ED431B 2017/01). We gratefully acknowledge NVIDIA Corporation for the donation of a GTX Titan X GPU., Xunta de Galicia; ED431B 2017/01
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

A Transition-Based Algorithm for Unrestricted AMR Parsing

  • Vilares, David
  • Gómez-Rodríguez, Carlos
The Conference was held in New Orleans, June 1 to June 6, 2018 at the Hyatt Regency New Orleans hotel., [Absctract]: Non-projective parsing can be useful to handle cycles and reentrancy in AMR graphs. We explore this idea and introduce a greedy left-to-right non-projective transition-based parser. At each parsing configuration, an oracle decides whether to create a concept or whether to connect a pair of existing concepts. The algorithm handles reentrancy and arbitrary cycles natively, i.e. within the transition system itself. The model is evaluated on the LDC2015E86 corpus, obtaining results close to the state of the art, including a Smatch of 64%, and showing good behavior on reentrant edges., This work is funded from the European Research Council (ERC), under the European Union’s Horizon 2020 research and innovation programme (FASTPARSE, grant agreement No 714150), from the TELEPARES-UDC project (FFI2014- 51978-C2-2-R) and the ANSWER-ASAP project (TIN2017-85160-C2-1-R) from MINECO, and from Xunta de Galicia (ED431B 2017/01). We gratefully acknowledge NVIDIA Corporation for the donation of a GTX Titan X GPU., Xunta de Galicia; ED431B 2017/01
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Syntactically enriched multilingual sentiment analysis

  • Gómez-Rodríguez, Carlos
Proceedings of the 2020 Workshop on Hybrid Intelligence for Natural Language Processing Tasks, HI4NLP (co-located at ECAI-2020), Santiago de Compostela, August 29, 2020, published at http://ceur-ws.org, Copyright © 2020 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0)., [Abstract]: Sentiment analysis of natural language texts needs to deal with linguistic phenomena like negation, intensification or adversative clauses. In this talk, I present an approach to tackle such phenomena by means of syntactic information. Our approach combines machine learning and symbolic processing: the former is used to obtain dependency trees for input sentences, and the latter to obtain the sentiment polarity for each sentence using handwritten rules that traverse the tree. Thanks to universal guidelines for syntactic annotation, our approach is applicable to multiple languages without rewriting the rules. Additionally, very accurate parsing is not needed for our approach to be helpful: fast and simple parsers will do, even if they lag behind state-of-the-art accuracy., This work 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), from the ANSWER-ASAP project (TIN2017-85160-C2-1-R) from MINECO, and from Xunta de Galicia and ERDF (ED431B 2017/01, ED431G 2019/01)., Xunta de Galicia; ED431B 2017/01, Xunta de Galicia; ED431G 2019/01
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)

Towards Making a Dependency Parser See

  • Strzyz, Michalina
  • Vilares, David
  • Gómez-Rodríguez, Carlos
It was held in Hong Kong, China. 3-7 November, 2019, [Absctract]: We explore whether it is possible to leverage eye-tracking data in an RNN dependency parser (for English) when such information is only available during training - i.e. no aggregated or token-level gaze features are used at inference time. To do so, we train a multitask learning model that parses sentences as sequence labeling and leverages gaze features as auxiliary tasks. Our method also learns to train from disjoint datasets, i.e. it can be used to test whether already collected gaze features are useful to improve the performance on new non-gazed annotated treebanks. Accuracy gains are modest but positive, showing the feasibility of the approach. It can serve as a first step towards architectures that can better leverage eye-tracking data or other complementary information available only for training sentences, possibly leading to improvements in syntactic parsing., This work 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), from the ANSWER-ASAP project (TIN2017-85160-C2-1-R) from MINECO, and from Xunta de Galicia (ED431B 2017/01)., Xunta de Galicia; ED431B 2017/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Optimality of syntactic dependency distances

  • Ferrer-i-Cancho, Ramon
  • Gómez-Rodríguez, Carlos
  • Esteban, Juan Luis
  • Alemany-Puig, Lluís
[Abstract]: It is often stated that human languages, as other biological systems, are shaped by cost-cutting pressures but, to what extent? Attempts to quantify the degree of optimality of languages by means of an optimality score have been scarce and focused mostly on English. Here we recast the problem of the optimality of the word order of a sentence as an optimization problem on a spatial network where the vertices are words, arcs indicate syntactic dependencies, and the space is defined by the linear order of the words in the sentence. We introduce a score to quantify the cognitive pressure to reduce the distance between linked words in a sentence. The analysis of sentences from 93 languages representing 19 linguistic families reveals that half of languages are optimized to a 70 % or more. The score indicates that distances are not significantly reduced in a few languages and confirms two theoretical predictions: that longer sentences are more optimized and that distances are more likely to be longer than expected by chance in short sentences. We present a hierarchical ranking of languages by their degree of optimization. The score has implications for various fields of language research (dependency linguistics, typology, historical linguistics, clinical linguistics, and cognitive science). Finally, the principles behind the design of the score have implications for network science., We are grateful to C. Bentz, N. Català, M. H. Christiansen, M. Gustison, and A. Hernández-Fernández for helpful comments. R.F.C. and L.A.P. are supported by Grant No. TIN2017-89244-R from MINECO (Ministerio de Economía, Industria y Competitividad). R.F.C. is also supported by the recognition No. 2017SGR-856 (MACDA) from AGAUR (Generalitat de Catalunya). L.A.P. is also supported by Secretaria d’Universitats i Recerca de la Generalitat de Catalunya and the Social European Fund. C.G.R. is supported by the European Research Council (ERC), under the European Union’s Horizon 2020 research and innovation program (FASTPARSE, Grant No. 714150), the ANSWER-ASAP project (No. TIN2017-85160-C2-1-R) and SCANNER-UDC project (No. PID2020-113230RB-C21) from ERDF/MICINN-AEI, and Xunta de Galicia (ED431C 2020/11 and an Oportunius program grant to complement ERC grants); the CITIC research center is funded by ERDF and Xunta de Galicia (ERDFGalicia 2014-2020 program, Grant No. ED431G 2019/01). J.L.E. is funded by Grants No. TIN2016-76573-C2-1-P and No. PID2019-109137GB-C22 from MINECO. R.F.C. designed the research; R.F.C and C.G.R. wrote the paper; C.G.R. collected and preprocessed data; C.G.R. and R.F.C. analyzed the data; R.F.C, C.G.R., and J.L.E developed the mathematical arguments; C.G.R., J.L.E., and L.A.P. wrote the core code; all authors verified the mathematical arguments., Xunta de Galicia; ED431C 2020/11, Xunta de Galicia; ED431G 2019/01, Agència de Gestió d'Ajuts Universitaris i de Recerca; 2017SGR-856
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Parsing as Pretraining

  • Vilares, David
  • Strzyz, Michalina
  • Søgaard, Anders
  • Gómez-Rodríguez, Carlos
[Abstract] Recent analyses suggest that encoders pretrained for language modeling capture certain morpho-syntactic structure. However, probing frameworks for word vectors still do not report results on standard setups such as constituent and dependency parsing. This paper addresses this problem and does full parsing (on English) relying only on pretraining architectures – and no decoding. We first cast constituent and dependency parsing as sequence tagging. We then use a single feed-forward layer to directly map word vectors to labels that encode a linearized tree. This is used to: (i) see how far we can reach on syntax modelling with just pretrained encoders, and (ii) shed some light about the syntax-sensitivity of different word vectors (by freezing the weights of the pretraining network during training). For evaluation, we use bracketing F1-score and LAS, and analyze in-depth differences across representations for span lengths and dependency displacements. The overall results surpass existing sequence tagging parsers on the PTB (93.5%) and end-to-end EN-EWT UD (78.8%)., We thank Mark Anderson and Daniel Hershcovich for their comments. DV, MS and CGR are funded by the ERC under the European Union’s Horizon 2020 research and innovation programme (FASTPARSE, grant No 714150), by the ANSWER-ASAP project (TIN2017-85160-C2-1-R) from MINECO, and by Xunta de Galicia (ED431B 2017/01). AS is funded by a Google Focused Research Award, Xunta de Galicia; ED431B 2017/01
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Discontinuous Constituent Parsing as Sequence Labeling

  • Vilares, David
  • Gómez-Rodríguez, Carlos
EMNLP2020 took place online from November 16 – 20 2020, [Absctract]: This paper reduces discontinuous parsing to sequence labeling. It first shows that existing reductions for constituent parsing as labeling do not support discontinuities. Second, it fills this gap and proposes to encode tree discontinuities as nearly ordered permutations of the input sequence. Third, it studies whether such discontinuous representations are learnable. The experiments show that despite the architectural simplicity, under the right representation, the models are fast and accurate., We thank Maximin Coavoux for giving us access to the data used in this work. We acknowledge the European Research Council (ERC), which has funded this research under the European Union’s Horizon 2020 research and innovation programme (FASTPARSE, grant agreement No 714150), MINECO (ANSWER-ASAP, TIN2017-85160-C2- 1-R), Xunta de Galicia (ED431C 2020/11), and Centro de Investigacion de Galicia ”CITIC”, ´ funded by Xunta de Galicia and the European Union (European Regional Development FundGalicia 2014-2020 Program), by grant ED431G 2019/01., Xunta de Galicia; ED431C 2020/11, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Classification of the relative positions between a small ellipsoid and an elliptic paraboloid

  • Brozos-Vázquez, Miguel
  • Pereira-Sáez, María José
  • Souto Salorio, María José
  • Tarrío-Tobar, Ana D.
©2019. 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 Brozos-Vázquez, M., Pereira-Sáez, M. J., Souto-Salorio, M. J., & Tarrío-Tobar, A. D. (2019). “Classification of the relative positions between a small ellipsoid and an elliptic paraboloid” has been accepted for publication in Computer Aided Geometric Design, 72, 34–48. The Version of Record is available online at https://doi.org/10.1016/j.cagd.2019.05.002., [Abstract]: We classify all the relative positions between an ellipsoid and an elliptic paraboloid when the ellipsoid is small in comparison with the paraboloid (small meaning that the two surfaces cannot be tangent at two points simultaneously when one is moved with respect to the other). This provides an easy way to detect contact between the two surfaces by a direct analysis of the coefficients of a fourth degree polynomial., The authors wish to thank the referees for extremely valuable comments and suggestions, which were essential to improve the final version of the paper. Supported by Projects ED431F 2017/03, TIN2017-85160-C2-1-R, MTM2016-75897-P and MTM2016-78647-P (AEI/FEDER, UE)., Xunta de Galicia; ED431F 2017/03
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Discovering Topics in Twitter About the COVID-19 Outbreak in Spain, Descubriendo temas en Twitter sobre el brote del COVID-19 en España

  • Agüero-Torales, Marvin M.
  • Vilares, David
  • López-Herrera, Antonio G.
[Resumen] En este trabajo, analizamos lo que los usuarios han estado discutiendo en Twitter durante el comienzo de la pandemia causada por el COVID-19. Concretamente, analizamos tres fases diferenciadas de la crisis del COVID-19 en España: el propio tiempo de pre-crisis, el estallido de la enfermedad y el confinamiento. Para llevar esto a cabo, primero recolectamos una gran cantidad de tuits que son preprocesados. A continuación, agrupamos los tuits en distintas temáticas usando un modelo de Latent Dirichlet Allocation, y definimos estrategias generativas y discriminativas para extraer las palabras clave y oraciones más representativas para cada tema. Finalmente, incluimos un exhaustivo análisis cualitativo sobre dichos temas, y cómo estos se corresponden con distintas problemáticas surgidas en España en distintos momentos de la crisis., [Abstract] In this work, we apply topic modeling to study what users have been discussing in Twitter during the beginning of the COVID-19 pandemic. More particularly, we explore the period of time that includes three differentiated phases of the COVID-19 crisis in Spain: the pre-crisis time, the outbreak, and the beginning of the lockdown. To do so, we first collect a large corpus of Spanish tweets and clean them. Then, we cluster the tweets into topics using a Latent Dirichlet Allocation model, and define generative and discriminative routes to later extract the most relevant keywords and sentences for each topic. Finally, we provide an exhaustive qualitative analysis about how such topics correspond to the situation in Spain at different stages of the crisis., MMAT has been partially funded by Barcelona Supercomputing Center (BSC) through the Spanish Plan for advancement of Language Technologies `Plan TL' and the Secretaría de Estado de Digitalización e Inteligencia Artificial (SEDIA). DV is supported by MINECO (TIN2017-85160-C2-1-R), by Xunta de Galicia (ED431C 2020/11), by Centro de Investigación de Galicia `CITIC' (European Regional Development Fund-Galicia 2014-2020 Program, ED431G 2019/01), and by a 2020 Leonardo Grant for Researchers and Cultural Creators from the BBVA Foundation, Xunta de Galicia; ED431C 2020/11, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Multitask Pointer Network for Multi-Representational Parsing

  • Fernández-González, Daniel
  • Gómez-Rodríguez, Carlos
Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG, [Abstract] Dependency and constituent trees are widely used by many artificial intelligence applications for representing the syntactic structure of human languages. Typically, these structures are separately produced by either dependency or constituent parsers. In this article, we propose a transition-based approach that, by training a single model, can efficiently parse any input sentence with both constituent and dependency trees, supporting both continuous/projective and discontinuous/non-projective syntactic structures. To that end, we develop a Pointer Network architecture with two separate task-specific decoders and a common encoder, and follow a multitask learning strategy to jointly train them. The resulting quadratic system, not only becomes the first parser that can jointly produce both unrestricted constituent and dependency trees from a single model, but also proves that both syntactic formalisms can benefit from each other during training, achieving state-of-the-art accuracies in several widely-used benchmarks such as the continuous English and Chinese Penn Treebanks, as well as the discontinuous German NEGRA and TIGER datasets., We acknowledge the European Research Council (ERC), which has funded this research under the European Union’s Horizon 2020 research and innovation programme (FASTPARSE, grant agreement No 714150), ERDF/MICINN-AEI (ANSWER-ASAP, TIN2017-85160-C2-1-R; SCANNER-UDC, PID2020-113230RB-C21), Xunta de Galicia, Spain (ED431C 2020/11), and Centro de Investigación de Galicia “CITIC”, funded by Xunta de Galicia, Spain and the European Union (ERDF - Galicia 2014–2020 Program), by grant ED431G 2019/01. Funding for open access charge: Universidade da Coruña / CISUG, Xunta de Galicia; ED431C 2020/11, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

HEAD-QA: A Healthcare Dataset for Complex Reasoning

  • Vilares, David
  • Gómez-Rodríguez, Carlos
Took place in Florence (Italy) at the 'Fortezza da Basso' from July 28th to August 2nd, 2019., [Absctract]: We present HEAD-QA, a multi-choice question answering testbed to encourage research on complex reasoning. The questions come from exams to access a specialized position in the Spanish healthcare system, and are challenging even for highly specialized humans. We then consider monolingual (Spanish) and cross-lingual (to English) experiments with information retrieval and neural techniques. We show that: (i) HEAD-QA challenges current methods, and (ii) the results lag well behind human performance, demonstrating its usefulness as a benchmark for future work., This work has received support from the TELEPARES-UDC project (FFI2014-51978-C2- 2-R) and the ANSWER-ASAP project (TIN2017- 85160-C2-1-R) from MINECO, from Xunta de Galicia (ED431B 2017/01), and from the European Research Council (ERC), under the European Union’s Horizon 2020 research and innovation programme (FASTPARSE, grant agreement No 714150). We thank Mark Anderson for his help with translation fluency evaluation., Xunta de Galicia; ED431B 2017/01
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Transition-based Semantic Dependency Parsing with Pointer Networks

  • Fernández-González, Daniel
  • Gómez-Rodríguez, Carlos
[Abstract]: Transition-based parsers implemented with Pointer Networks have become the new state of the art in dependency parsing, excelling in producing labelled syntactic trees and outperforming graph-based models in this task. In order to further test the capabilities of these powerful neural networks on a harder NLP problem, we propose a transition system that, thanks to Pointer Networks, can straightforwardly produce labelled directed acyclic graphs and perform semantic dependency parsing. In addition, we enhance our approach with deep contextualized word embeddings extracted from BERT. The resulting system not only outperforms all existing transition-based models, but also matches the best fully-supervised accuracy to date on the SemEval 2015 Task 18 datasets among previous state-of-the-art graph-based parsers., This work 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), from the ANSWER-ASAP project (TIN2017-85160-C2-1-R) from MINECO, and from Xunta de Galicia (ED431B 2017/01, ED431G 2019/01)., Xunta de Galicia; ED431B 2017/01, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Faster shift-reduce constituent parsing with a non-binary, bottom-up strategy

  • Fernández-González, Daniel
  • Gómez-Rodríguez, Carlos
© 2019. This is the final peer-reviewed manuscript that was accepted for publication at Artificial Intelligence and made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/ licenses/by-nc-nd/4.0/. This may not reflect subsequent changes resulting from the publishing process such as editing, formatting, pagination, and other quality control mechanisms. The final journal publication is available at https://doi.org/10.1016/j.artint.2019.07.006, [Absctract]: An increasingly wide range of artificial intelligence applications rely on syntactic information to process and extract meaning from natural language text or speech, with constituent trees being one of the most widely used syntactic formalisms. To produce these phrase-structure representations from sentences in natural language, shift-reduce constituent parsers have become one of the most efficient approaches. Increasing their accuracy and speed is still one of the main objectives pursued by the research community so that artificial intelligence applications that make use of parsing outputs, such as machine translation or voice assistant services, can improve their performance. With this goal in mind, we propose in this article a novel non-binary shift-reduce algorithm for constituent parsing. Our parser follows a classical bottom-up strategy but, unlike others, it straightforwardly creates non-binary branchings with just one transition, instead of requiring prior binarization or a sequence of binary transitions, allowing its direct application to any language without the need of further resources such as percolation tables. As a result, it uses fewer transitions per sentence than existing transition-based constituent parsers, becoming the fastest such system and, as a consequence, speeding up downstream applications. Using static oracle training and greedy search, the accuracy of this novel approach is on par with state-of-the-art transition-based constituent parsers and outperforms all top-down and bottom-up greedy shift-reduce systems on the Wall Street Journal section from the English Penn Treebank and the Penn Chinese Treebank. Additionally, we develop a dynamic oracle for training the proposed transition-based algorithm, achieving further improvements in both benchmarks and obtaining the best accuracy to date on the Penn Chinese Treebank among greedy shift-reduce parsers., This work 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), from the ANSWER-ASAP project (TIN2017-85160-C2-1-R) from MINECO, and from Xunta de Galicia (ED431B 2017/01)., Xunta de Galicia; ED431B 2017/01
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Better, Faster, Stronger Sequence Tagging Constituent Parsers

  • Vilares, David
  • Abdou, Mostafa
  • Søgaard, Anders
NAACL-HLT 2019 was held at the Hyatt Regency in Minneapolis from June 2nd through 7th, 2019., [Absctract]: Sequence tagging models for constituent parsing are faster, but less accurate than other types of parsers. In this work, we address the following weaknesses of such constituent parsers: (a) high error rates around closing brackets of long constituents, (b) large label sets, leading to sparsity, and (c) error propagation arising from greedy decoding. To effectively close brackets, we train a model that learns to switch between tagging schemes. To reduce sparsity, we decompose the label set and use multi-task learning to jointly learn to predict sublabels. Finally, we mitigate issues from greedy decoding through auxiliary losses and sentence-level fine-tuning with policy gradient. Combining these techniques, we clearly surpass the performance of sequence tagging constituent parsers on the English and Chinese Penn Treebanks, and reduce their parsing time even further. On the SPMRL datasets, we observe even greater improvements across the board, including a new state of the art on Basque, Hebrew, Polish and Swedish., DV has received support from the European Research Council (ERC), under the European Union’s Horizon 2020 research and innovation programme (FASTPARSE, grant agreement No 714150), from the TELEPARES-UDC project (FFI2014-51978-C2-2-R) and the ANSWERASAP project (TIN2017-85160-C2-1-R) from MINECO, and from Xunta de Galicia (ED431B 2017/01). MA and AS are funded by a Google Focused Research Award., Xunta de Galicia; ED431B 2017/01
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Constituent Parsing as Sequence Labeling

  • Gómez-Rodríguez, Carlos
  • Vilares, David
EMNLP 2018, Square Meeting Center, Brussels. From October 31st through November 4th., [Absctract]: We introduce a method to reduce constituent parsing to sequence labeling. For each word wt, it generates a label that encodes: (1) the number of ancestors in the tree that the words wt and wt+1 have in common, and (2) the nonterminal symbol at the lowest common ancestor. We first prove that the proposed encoding function is injective for any tree without unary branches. In practice, the approach is made extensible to all constituency trees by collapsing unary branches. We then use the PTB and CTB treebanks as testbeds and propose a set of fast baselines. We achieve 90% F-score on the PTB test set, outperforming the Vinyals et al. (2015) sequence-to-sequence parser. In addition, sacrificing some accuracy, our approach achieves the fastest constituent parsing speeds reported to date on PTB by a wide margin., This work 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), from the TELEPARESUDC project (FFI2014-51978-C2-2-R) and the ANSWER-ASAP project (TIN2017-85160-C2-1-R) from MINECO, and from Xunta de Galicia (ED431B 2017/01). We gratefully acknowledge NVIDIA Corporation for the donation of a GTX Titan X GPU., Xunta de Galicia; ED431B 2017/01
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

A Unifying Theory of Transition-based and Sequence Labeling Parsing

  • Gómez-Rodríguez, Carlos
  • Strzyz, Michalina
  • Vilares, David
Held online due to COVID-19. December 2020, Barcelona, Spain, [Absctract]: We define a mapping from transition-based parsing algorithms that read sentences from left to right to sequence labeling encodings of syntactic trees. This not only establishes a theoretical relation between transition-based parsing and sequence-labeling parsing, but also provides a method to obtain new encodings for fast and simple sequence labeling parsing from the many existing transition-based parsers for different formalisms. Applying it to dependency parsing, we implement sequence labeling versions of four algorithms, showing that they are learnable and obtain comparable performance to existing encodings., This work has received funding from the European Research Council (ERC), which has funded this research under the European Union’s Horizon 2020 research and innovation programme (FASTPARSE, grant agreement No 714150), from MINECO (ANSWER-ASAP, TIN2017-85160-C2-1-R), from Xunta de Galicia (ED431C 2020/11), and from Centro de Investigacion de Galicia ‘CITIC’, funded by Xunta de ´ Galicia and the European Union (European Regional Development Fund- Galicia 2014-2020 Program), by grant ED431G 2019/01. DV is supported by a 2020 Leonardo Grant for Researchers and Cultural Creators from the BBVA Foundation., Xunta de Galicia; ED431C 2020/11, Xunta de Galicia; ED431G 2019/01