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): 23
Encontrada(s) 1 página(s)

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
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 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 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 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 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 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 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 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