Search

This paper describes the submission to the WMT 2021 news translation shared task by the UPC Machine Translation group. The goal of the task is to translate German to French (De-Fr) and French to German (Fr-De). Our submission focuses on fine-tuning a pre-trained model to take advantage of monolingual data. We fine-tune mBART50 using the filtered data, and additionally, we train a Transformer model on the same data from scratch. In the experiments, we show that fine-tuning mBART50 results in 31.69 BLEU for De-Fr and 23.63 BLEU for Fr-De, which increases 2.71 and 1.90 BLEU accordingly, as compared to the model we train from scratch. Our final submission is an ensemble of these two models, further increasing 0.3 BLEU for Fr-De.

State of the art neural network approaches enable massive multilingual translation. How close are we to universal translation in any language?

State-of-the-art multilingual machine translation relies on a universal encoder-decoder, which requires retraining the entire system to add new languages. In this paper, we propose an alternative approach that is based on language-specific encoder-decoders, and can thus be more easily extended to new languages by learning their corresponding modules. So as to encourage a common interlingua representation, we simultaneously train the N initial languages. Our experiments show that the proposed approach outperforms the universal encoder-decoder by 3.28 BLEU points on average, while allowing to add new languages without the need to retrain the rest of the modules. All in all, our work closes the gap between shared and language-specific encoderdecoders, advancing toward modular multilingual machine translation systems that can be flexibly extended in lifelong learning settings. This work is supported by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No. 947657).

This paper reports the optimization of using the out-of-domain data in the Biomedical translation task. We firstly optimized our parallel training dataset using the BabelNet in-domain terminology words. Afterward, to increase the training set, we studied the effects of the out-of-domain data on biomedical translation tasks, and we created a mixture of in-domain and out-of-domain training sets and added more in-domain data using forward translation in the English-Spanish task. Finally, with a simple bpe optimization method, we increased the number of in-domain subwords in our mixed training set and trained the Transformer model on the generated data. Results show improvements using our proposed method. © 2021 Association for Computational Linguistics This work was supported by the project ADAVOICE, PID2019-107579RB-I00 / AEI 256 / 10.13039/501100011033.

A lifelong learning system can adapt to new data without forgetting previously acquired knowledge. In this paper, we introduce the first benchmark for lifelong learning machine translation. For this purpose, we provide training, lifelong and test data sets for two language pairs: English-German and English-French. Additionally, we report the results of our baseline systems, which we make available to the public. The goal of this shared task is to encourage research on the emerging topic of lifelong learning machine translation. This work is is supported in part by the Spanish Ministerio de Ciencia e Innovacion, through the postdoctoral senior grant Ramon y Cajal and by the Agencia Estatal de Investigacion through the projects EUR2019-103819, PCIN-2017-079 and PID2019-107579RB-I00 / AEI / 10.13039/501100011033.

La intel·ligència artificial mou milions d’euros i ocupa una part important de les agendes polítiques i estratègiques dels governs. En aquest article reflexionem sobre aquest concepte difús i ho fem a través de tres entrevistes a tres persones que ocupen llocs rellevants en aquest àmbit, i que a més tenen en comú que són nascudes i formades a Catalunya. Amb elles hem parlat distesament dels orígens d’aquesta revolució tecnològica, del seu impacte sobre l’organització social i el mercat laboral, de les seves implicacions ètiques, de les tendències futures que es dibuixen i també sobre quin paper juguem els catalans en aquesta revolució. En la qüestió catalana hem volgut saber si les nostres universitats són capaces de generar el talent necessari, com impulsar estratègicament el teixit empresarial en aquesta àrea, quin paper ha de tenir l’administració i què hauria de passar perquè la intel·ligència artificial també parli en català. En aquestes converses, els nostres interlocutors ens han transmès la urgència i la necessitat de treballar en la definició d’estratègies concretes de país per tal de no quedar socialment i estructuralment enrere en el que sembla una revolució sense límits Artificial intelligence moves millions of euros and occupies an important part of governments’ political and strategic agendas. In this article we reflect on this varied concept through interviews with three people who hold important positions in the field of artificial intelligence and who also have in common the fact that they were born and trained in Catalonia. We talked with them about the origins of this technological revolution, its impact on social organization and the labour market, its ethical implications, the future trends which are emerging and also on what role we Catalans play in this. With regards to the Catalan question we wanted to know if our universities are capable of generating the necessary talent, how to strategically boost the fabric of business in this area, what role the Administration should play and what should happen for artificial intelligence to communicate in Catalan. Throughout these conversations, the interviewees have conveyed to us the urgency and need to work on the definition of country-specific strategies in order not to be left behind socially and structurally in what seems to be a revolution without limits. Aquest treball ha estat finançat en part pel Ministerio de Ciencia e Innovación mitjançant el programa Ramón i Cajal i per l’European Research Council (ERC) sota European Union’s Horizon 2020 research and innovation programme (grant agreement No 947657).

In this paper, we present two productive and functional recommender methods to improve the accuracy of predicting the right product for the user. One proposal is a survey-based recommender system that uses k-nearest neighbors. It recommends products by asking questions from the user, efficiently applying a binary product vector to the product attributes, and processing the request with a minimum error. The second proposal uses an enriched collaborative-based recommender system using enriched weighted vectors. Thanks to the style rules, the enriched collaborative based method recommends outfits with competitive recommendation quality. We evaluated both of the proposals on a Kaggle fashion-dataset along with iMaterialist and, results show equivalent performance on binary gender and product attributes.

Cross-lingual word embeddings aim to bridge the gap between high-resource and low-resource languages by allowing to learn multilingual word representations even without using any direct bilingual signal. The lion's share of the methods are projection-based approaches that map pre-trained embeddings into a shared latent space. These methods are mostly based on the orthogonal transformation, which assumes language vector spaces to be isomorphic. However, this criterion does not necessarily hold, especially for morphologically-rich languages. In this paper, we propose a self-supervised method to refine the alignment of unsupervised bilingual word embeddings. The proposed model moves vectors of words and their corresponding translations closer to each other as well as enforces length- and center-invariance, thus allowing to better align cross-lingual embeddings. The experimental results demonstrate the effectiveness of our approach, as in most cases it outperforms state-of-the-art methods in a bilingual lexicon induction task. We thank anonymous reviewers for their helpful comments. This work is supported in part by the Spanish Ministerio de Economía y Competitividad, the European Regional Development Fund and the Agencia Estatal de Investigación, through the post-doctoral senior grant Ramón y Cajal, the contract TEC2015-69266-P (MINECO/FEDER,EU) and the contract PCIN-2017-079 (AEI/MINECO).

Misrepresentation of certain communities in datasets is causing big disruptions in artificial intelligence applications. In this paper, we propose using an automatically extracted gender-balanced dataset parallel corpus from Wikipedia. This balanced set is used to perform fine-tuning techniques from a bigger model trained on unbalanced datasets to mitigate gender biases in neural machine translation. This work is supported in part by the Catalan Agency for Management of University and Research Grants (AGAUR) through an Industrial Ph.D. Grant, the Spanish Ministerio de Ciencia e Innovacion and by the Agencia Estatal de Investigacion, through the postdoctoral senior grant Ramón y Cajal and by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No. 947657).

In this paper, we describe the TALP-UPC participation in the WMT Similar Language Translation task between Catalan, Spanish, and Portuguese, all of them, Romance languages. We made use of different techniques to improve the translation between these languages. The multilingual shared encoder/decoder has been used for all of them. Additionally, we applied back-translation to take advantage of the monolingual data. Finally, we have applied fine-tuning to improve the in-domain data. Each of these techniques brings improvements over the previous one. In the official evaluation, our system was ranked 1st in the Portuguese-to-Spanish direction, 2nd in the opposite direction, and 3rd in the Catalan-Spanish pair. This work is supported in part by the Spanish Ministerio de Ciencia e Innovacion, through the postdoctoral senior grant Ramon y Cajal and by the Agencia Estatal de Investigacion through the projects EUR2019-103819, PCIN2017-079 and PID2019-107579RB-I00 / AEI / 10.13039/501100011033

In this article, we describe the TALP-UPC participation in the WMT20 news translation shared task for Tamil-English. Given the low amount of parallel training data, we resort to adapt the task to a multilingual system to benefit from the positive transfer from high resource languages. We use iterative backtranslation to fine-tune the system and benefit from the monolingual data available. In order to measure the effectivity of such methods, we compare our results to a bilingual baseline system. This work is supported in part by the Google Faculty Research Award 2019, the Spanish Ministerio de Ciencia e Innovacion, through the postdoctoral senior grant Ramon y Cajal and by the Agencia Estatal de Investigacion through the projects EUR2019-103819, PCIN-2017-079 and PID2019- 107579RB-I00 / AEI / 10.13039/501100011033

El objetivo del proyecto AMALEU es aprender una representación común para diferentes idiomas. Se pretende tener una representación común para la lengua oral y una para la lengua escrita. AMALEU, de dos años de duración, está financiado por el MINECO dentro del programa de Europa Excelencia. The objective of AMALEU’s project is learning a multilingual common representation for speech and a different multilingual common representation for text. AMALEU is a two-year project funded by the MINECO. Este trabajo está financiado por el Ministerio de Economía y Competitividad a través del projecto EUR2019-103819 y el programa Ramón y Cajal que incluye financiamiento del European Regional Development Fund.

The main alternatives nowadays to dealwith sequences are Recurrent Neural Net-works (RNN), Convolutional Neural Networks(CNN) architectures and the Transformer. Inthis context, RNN’s, CNN’s and Transformerhave most commonly been used as an encoder-decoder architecture with multiple layers ineach module. Far beyond this, these architec-tures are the basis for the contextual word em-beddings which are revolutionizing most natural language downstream applications. However, intermediate layer representations insequence-based architectures can be difficultto interpret. To make each layer representation within these architectures more accessible and meaningful, we introduce a web-based toolthat visualizes them both at the sentence and token level. We present three use cases. The first analyses gender issues in contextual worde mbeddings. The second and third are show-ing multilingual intermediate representations for sentences and tokens and the evolution of these intermediate representations along the multiple layers of the decoder and in the con-text of multilingual machine translation. This work is supported by a Google Faculty Research Award. This workis also supported by the Spanish Ministerio de Economía y Competitividad, the European Regional Development Fund and the Agencia Estatal de Investigación, through the post-doctoral senior grant Ramón y Cajal, contracts TEC2015-69266-P and TIN2017-88515-C2-1-R(GEN3DLIVE) (MINECO/FEDER,EU), and contract PCIN-2017-079 (AEI/MINECO).

n the last years, deep learning algorithms have highly revolutionized several areas including speech, image and natural language processing. The specific field of Machine Translation (MT) has not remained invariant. Integration of deep learning in MT varies from re-modeling existing features into standard statistical systems to the development of a new architecture. Among the different neural networks, research works use feed- forward neural networks, recurrent neural networks and the encoder-decoder schema. These architectures are able to tackle challenges as having low-resources or morphology variations. This manuscript focuses on describing how these neural networks have been integrated to enhance different aspects and models from statistical MT, including language modeling, word alignment, translation, reordering, and rescoring. Then, we report the new neural MT approach together with a description of the foundational related works and recent approaches on using subword, characters and training with multilingual languages, among others. Finally, we include an analysis of the corresponding challenges and future work in using deep learning in MT

In this article we describe the TALP-UPC research group participation in the WMT18 news shared translation task for FinnishEnglish and Estonian-English within the multi-lingual subtrack. All of our primary submissions implement an attention-based Neural Machine Translation architecture. Given that Finnish and Estonian belong to the same language family and are similar, we use as training data the combination of the datasets of both language pairs to paliate the data scarceness of each individual pair. We also report the translation quality of systems trained on individual language pair data to serve as baseline and comparison reference.

This paper presents experiments compar- ing character-based and byte-based neural machine translation systems. The main motivation of the byte-based neural ma- chine translation system is to build multi- lingual neural machine translation systems that can share the same vocabulary. We compare the performance of both systems in several language pairs and we see that the performance in test is similar for most language pairs while the training time is slightly reduced in the case of byte-based neural machine translation.

Phrase-based machine translation assumes that all words are at the same distance and translates them using feature functions that approximate the probability at different levels. On the other hand, neural machine translation infers a word embedding and translates these word vectors using a neural model. At the moment, both approaches co-exist and are being intensively investigated. This paper to the best of our knowledge is the first work that both compares and combines these two systems by: using the phrase-based output to solve unknown words in the neural machine translation output; using the neural alignment in the phrase-based system; comparing how the popular strategy of pre-reordering aspects both systems; and combining both translation outputs. Improvements are achieved in Catalan-to- Spanish and German-to-English.

La variación morfológica entre un lenguaje fuente y el lenguaje destino genera dificultades a los algoritmos estándares de traducción como el estadístico basado en segmentos. En este trabajo planteamos dividir la tarea de traducción en dos partes: primero, simplificamos el lenguaje destino en términos morfológicos y construimos el sistema de traducción con esta modificación; y después utilizamos un algoritmo de clasificación para generar la morfología final. Este trabajo presenta una arquitectura de aprendizaje profundo que permite añadir de manera efectiva la información morfológica a la traducción simplificada generada por un traductor estadístico basado en segmentos. Demostramos que la arquitectura diseñada presenta resultados superiores a los algoritmos estado-del-arte en términos de precisión y que la calidad de la traducción mejora en términos de METEOR

In recent years, Neural Machine Translation (NMT) has achieved state-of-the-art performance in translating from a language; source language, to another; target language. However, many of the proposed methods use word embedding techniques to represent a sentence in the source or target language. Character embedding techniques for this task has been suggested to represent the words in a sentence better. Moreover, recent NMT models use attention mechanism where the most relevant words in a source sentence are used to generate a target word. The problem with this approach is that while some words are translated multiple times, some other words are not translated. To address this problem, coverage model has been integrated into NMT to keep track of already-translated words and focus on the untranslated ones. In this research, we present a new architecture in which we use character embedding for representing the source and target languages, and also use coverage model to make certain that all words are translated. Experiments were performed to compare our model with coverage and character model and the results show that our model performs better than the other two models.

Este proyecto propone el desarrollo de nuevas arquitecturas para el procesado de la voz y el audio mediante métodos de aprendizaje profundo, explorando también nuevas aplicaciones y dando continuidad al trabajo inicial del equipo de investigadores solicitante y de toda la comunidad internacional. Las lineas de investigación incluyen: reconocimiento de voz, reconocimiento de eventos acústicos, síntesis de voz y traducción automática

There has been some exciting major progress in recent years in data analysis methods, including a variety of deep learning architectures, as well as further advances in kernel-based learning methods, which have demonstrated predictive superiority. In this paper we provide a brief motivated survey of recent proposals to explicitly or implicitly combine kernel methods with the notion of deep learning networks.

This paper describes the TALP–UPC system in the Spanish–English WMT 2016 biomedical shared task. Our system is a standard phrase-based system enhanced with vocabulary expansion using bilingual word embeddings and a characterbased neural language model with rescoring. The former focuses on resolving outof- vocabulary words, while the latter enhances the fluency of the system. The two modules progressively improve the final translation as measured by a combination of several lexical metrics.

Bidirectional Recurrent Neural Networks (BiRNNs) have shown outstanding results on sequence-to-sequence learning tasks. This architecture becomes specially interesting for multimodal machine translation task, since BiRNNs can deal with images and text. On most translation systems the same word embedding is fed to both BiRNN units. In this paper, we present several experiments to enhance a baseline sequence-to-sequence system (Elliott et al., 2015), for example, by using double embeddings. These embeddings are trained on the forward and backward direction of the input sequence. Our system is trained, validated and tested on the Multi30K dataset (Elliott et al., 2016) in the context of theWMT 2016Multimodal Translation Task. The obtained results show that thedouble-embedding approach performs significantly better than the traditional single-embedding one.

La Traducción Automática (TA) es un campo altamente interdisciplinar y multidisciplinar porque en él trabajan: ingenieros, informáticos, estadísticos y lingüistas. El objetivo de este proyecto es acercar los diferentes perfiles de la comunidad de la TA para plantear un paradigma integrado de TA que incluya tecnologías lingüísticas y estadísticas. Básicamente, nuestra investigación se centra en el problema de integrar dinámicamente dos de los paradigmas de traducción más populares: el basado en reglas y el estadístico. Una de las principales ideas es usar tecnologías lingüísticas desarrolladas para los sistemas basados en reglas o en el contexto del procesamiento del lenguaje natural. El nuevo paradigma proporcionará soluciones a los retos actuales de la TA como palabras desconocidas, reordenamiento y ambigüedades semánticas. El proyecto se focaliza en tres de las lenguas más hablados en el mundo: Chino, Castellano e Inglés; y todas las combinaciones de traducción entre ellas. Estos pares de lenguas no solo involucran intereses económicos y culturales, sino que además tienen importantes retos de TA como el morfológico, sintáctico y semántico

With the increasingly global nature of our everyday interactions, the need for multilingual technologies to support efficient and efective information access and communication cannot be overemphasized. Computational modeling of language has been the focus of Natural Language Processing, a subdiscipline of Artificial Intelligence. One of the current challenges for this discipline is to design methodologies and algorithms that are cross-language in order to create multilingual technologies rapidly. The goal of this JAIR special issue on Cross-Language Algorithms and Applications (CLAA) is to present leading research in this area, with emphasis on developing unifying themes that could lead to the development of the science of multi- and cross-lingualism. In this introduction, we provide the reader with the motivation for this special issue and summarize the contributions of the papers that have been included. The selected papers cover a broad range of cross-lingual technologies including machine translation, domain and language adaptation for sentiment analysis, cross-language lexical resources, dependency parsing, information retrieval and knowledge representation. We anticipate that this special issue will serve as an invaluable resource for researchers interested in topics of cross-lingual natural language processing.

Transliteration is the phonetic translation between two different languages. There are many works that approach transliteration using machine translation methods. This paper describes the official baseline system for the NEWS 2016 workshop shared task. This baseline is based on a standard phrase-based machine translation system using Moses. Results are between the range of best and worst from last year’s workshops providing a nice starting point for participants this year.

Recently, translation systems based on neural networks are starting to compete with systems based on phrases. The systems which are based on neural networks use vectorial repre- sentations of words. However, one of the biggest challenges that machine translation still faces, is dealing with large vocabularies and morphologically rich languages. This work aims to adapt a neural machine translation system to translate from Chinese to Spanish, using as input different types of granularity: words, characters, bitmap fonts of Chinese characters or words. The fact of performing the interpretation of every character or word as a bitmap font allows for obtaining more informed vectorial representations. Best results are obtained when using the information of the word bitmap font.

Neural Machine Translation (MT) has reached state-of-the-art results. However, one of the main challenges that neural MT still faces is dealing with very large vocabularies and morphologically rich languages. In this paper, we propose a neural MT system using character-based embeddings in combination with convolutional and highway layers to replace the standard lookup-based word representations. The resulting unlimited-vocabulary and affixaware source word embeddings are tested in a state-of-the-art neural MT based on an attention-based bidirectional recurrent neural network. The proposed MT scheme provides improved results even when the source language is not morphologically rich. Improvements up to 3 BLEU points are obtained in the German-English WMT task.

La traducción automática ha recibido mucho interés en el campo del procesado del lenguaje natural porque es un tema de interés social. Al mismo tiempo se trata de un problema interesante a nivel académico porque engloba diferentes tareas del tratamiento del lenguaje textual como la desambiguación léxica, el parseado o el reconocimiento de entidades. Este artículo presenta los últimos avances en esta área para las dos lenguas que encabezan el ranking de número de habitantes nativos: el chino y español. La investigación en traducción automática para ambos incluye aproximaciones basadas en reglas como estadísticas. El hecho que ambas aproximaciones estén activas deja el camino abierto para hacer una aproximación híbrida. De este modo, utilizando el caso particular del chino y español, este artículo: (1) describe las motivaciones económicas, sociales y académicas de este para ambos idiomas; (2) revisa, describe y muestra experimentos de las dos aproximaciones más populares de traducción automática (basada en reglas y estadística); (3) y dibuja líneas futuras que están siendo muy populares en el campo como es la aproximación híbrida.

Chinese and Spanish have different morphology structures, which poses a big challenge for translating between this pair of languages. In this paper, we analyze several strategies to better generalize from the Chinese non-morphology-based language to the Spanish rich morphologybased language. Strategies use a first-step of Spanish morphology-based simplifications and a second-step of fullform generation. The latter can be done using a translation system or classification methods. Finally, both steps are combined either by concatenation in cascade or integration using a factored-based style. Ongoing experiments (based on the United Nations corpus) and their results are described.

A comparison of interlingua and query translation is proposed in a particular cross-language information retrieval (CLIR) application which consists on retrieving a book from the collection by using one of its chapters in a different language as a query. The experiments are performed in three languages (English, Chinese and Spanish) and all the possible combinations. It is shown that interlingua is able to outperform the query translation approach in some cross-language tasks. Results are further analysed and it is found that, for this particular task, the quality of translation (in terms of BLEU and PER) is not directly correlated with the query translation performance.

Assessment in education allows for obtaining, organizing, and presenting information about how much and how well the student is learning. The current paper aims at analysing and discussing some of the most state-of-the-art assessment systems in education. Later, this work presents a specific use case developed for the Universitat Oberta de Catalunya, which is an online university. An automatic evaluation tool is proposed that allows the student to evaluate himself anytime and receive instant feedback. This tool is a web-based platform, and it has been designed for engineering subjects (i.e., with math symbols and formulas) in Catalan and Spanish. Particularly, the technique used for automatic assessment is latent semantic analysis. Although the experimental framework from the use case is quite challenging, results are promising.

Evaluation of machine translation output is an important task. Various human evaluation techniques as well as automatic metrics have been proposed and investigated in the last decade. However, very few evaluation methods take linguistic aspect into account. In this paper, we use an objective evaluation method for machine translation output that classifies all translation errors into one of the five following linguistic levels: orthographic, morphological, lexical, semantic and syntactic, in order to analyse its linguistic quality. Linguistic guidelines for the target language are required, and human evaluators use them in to classify the output errors. The experiments are performed on English-to-Catalan and Spanish-to-Catalan translation outputs generated by four different systems: two rule-based and two statistical. All translations are evaluated using three following methods: a standard human evaluation method, several widely used automatic metrics, and the linguistic evaluation. Pearson and Spearman correlation coefficients between the linguistic, human and automatic results are then calculated, showing that the semantic level correlates significantly with both human evaluation and automatic metrics. This work has been partially funded by the Spanish Department of Science and Innovation through the Juan de la Cierva fellowship program and the BUCEADOR project (TEC2009-14094-C04-01). The authors also want to thank the Barcelona Media Innovation Centre for its support and permission to publish this research. This work has been partially funded by the Spanish Department of Science and Innovation through the Juan de la Cierva fellowship program and the BUCEADOR project (TEC2009-14094-C04-01). The authors also want to thank the Barcelona Media Innovation Centre for its support and permission to publish this research. This work has been partially funded by the Spanish Department of Science and Innovation through the Juan de la Cierva fellowship program and the BUCEADOR project (TEC2009-14094-C04-01). The authors also want to thank the Barcelona Media Innovation Centre for its support and permission to publish this research.

This paper proposes to introduce a novel reordering model in the open-source Moses toolkit. The main idea is to provide weighted reordering hypotheses to the SMT decoder. These hypotheses are built using a first-step Ngram-based SMT translation from a source language into a third representation that is called reordered source language. Each hypothesis has its own weight provided by the Ngram-based decoder. This proposed reordering technique offers a better and more efficient translation when compared to both the distance-based and the lexicalized reordering. In addition to this reordering approach, this paper describes a domain adaptation technique which is based on a linear combination of an specific indomain and an extra out-domain translation models. Results for both approaches are reported in the Arabic-to-English 2008 IWSLT task. When implementing the weighted reordering hypotheses and the domain adaptation technique in the final translation system, translation results reach improvements up to 2.5 BLEU compared to a standard state-of-the-art Moses baseline system.

This paper describes the Barcelona Media Innovation Center participation in the 2nd International Competition on Plagiarism Detection. Particularly, our system focused on the external plagiarism detection task, which assumes the source documents are available. We present a two-step a approach. In the first step of our method, we build an information retrieval system based on Solr/Lucene, segmenting both suspicious and source documents into smaller texts.We perform a search based on bag-of-words which provides a first selection of potentially plagiarized texts. In the second step, each promising pair is further investigated. We implemented a sliding window approach that computes cosine distances between overlapping text segments from both the source and suspicious documents on a pair wise basis. As a result, a similarity matrix between text segments is obtained, which is smoothed by means of low-pass 2-D filtering. From the smoothed similarity matrix, plagiarized segments are identified by using image processing techniques. Our results were placed in the middle of the official ranking, which considered together two types of plagiarism: intrinsic and external.

Este trabajo presenta el uso de una técnica de recuperación de infor-mación crossling¨ue basada en escalamiento multidimensional para la iden-tificación de oraciones paralelas entre lenguas diferentes. El método prop-uesto permite hacer una reducción no-lineal del espacio de representación de las oraciones que se puede aprovechar para identificar similitudes semánti-cas entre conjuntos de oraciones en distintas lenguas. La técnica se ilustra con una colección pentaling¨ue extrai´da de la Constitución Española, la cual está disponible en las cuatro lenguas oficiales del estado español e inglés. Presentamos una evaluación comparativa entre nuestro método y un sistema de búsqueda crossling¨ue basado en la traducción automática de las consultas. Los resultados muestran que nuestro sistema mejora consistentemente en las 20 direcciones experimentales de búsqueda crossling¨ue que permite nuestra colección de datos

This study presents the TALP-UPC submission to the EACL Fourth Worskhop on Statistical Machine Translation 2009 evaluation campaign. It outlines the architecture and configuration of the 2009 phrase-based statistical machine translation (SMT) system, putting emphasis on the major novelty of this year: combination of SMT systems implementing different word reordering algorithms. Traditionally, we have concentrated on the Spanish-to-English and English-to-Spanish News Commentary translation tasks.

In this paper we propose a multilingual extension for OnAIR which is an ontology-aided information retrieval system applied to retrieve clips from a video collection. The multilingual extension basically involves allowing the user to search in several languages in a multilingual video collection. Particularly, the pair of languages we work in this paper are English and Portuguese. In order to perform query translation we use a statistical machine translation approach. Our experiments show that the multilingual system is capable of achieving almost the same quality of that obtained by the monolingual system.

This paper presents a web-based multimedia search engine built within the Buceador (www.buceador.org) research project. A proof-of-concept tool has been implemented which is able to retrieve information from a digital library made of multimedia documents in the 4 official languages in Spain (Spanish, Basque, Catalan and Galician). The retrieved documents are presented in the user language after translation and dubbing (the four previous languages + English). The paper presents the tool functionality, the architecture, the digital library and provide some information about the technology involved in the fields of automatic speech recognition, statistical machine translation, text-to-speech synthesis and information retrieval. Each technology has been adapted to the purposes of the presented tool as well as to interact with the rest of the technologies involved.

Chinese and Spanish are the most spoken languages in the world. However, there is not much research done in machine translation for this language pair. We experiment with the parallel Chinese-Spanish corpus (United Nations) to explore alternatives of SMT strategies which consist on using a pivot language. Particularly, two well-known alternatives are shown for pivoting: the cascade system and the pseudo-corpus. As Pivot language we use English, Arabic and French. Results show that English is the best pivot language between Chinese and Spanish. As a new strategy, we propose to perform a combination of the pivot strategies which is capable to highly outperform the direct translation strategy.

This paper describes an Information Retrieval engine that is used to support our Chinese-Portuguese machine translation services when no internet connection is available. Our mobile translation app, which is deployed on a portable device, relies by default on a server-based machine translation service, which is not accessible when no internet connection is available. For providing translation support under this condition, we have developed a contextualized off-line search engine that allows the users to continue using the app.

In this paper we present the first neural-based machine translation system trained to translate between standard national varieties of the same language. We take the pair Brazilian - European Portuguese as an example and compare the performance of this method to a phrase-based statistical machine translation system. We report a performance improvement of 0.9 BLEU points in translating from European to Brazilian Portuguese and 0.2 BLEU points when translating in the opposite direction. We also carried out a human evaluation experiment with native speakers of Brazilian Portuguese which indicates that humans prefer the output produced by the neural-based system in comparison to the statistical system., Comment: Proceedings of the Fifth Workshop on NLP for Similar Languages, Varieties and Dialects (VarDial)