Search

Computer-aided translation (CAT) tools based on translation memories (MT) play a prominent role in the translation workflow of professional translators. However, the reduced availability of in-domain TMs, as compared to in-domain monolingual corpora, limits its adoption for a number of translation tasks. In this paper, we introduce a novel neural approach aimed at overcoming this limitation by exploiting not only TMs, but also in-domain target-language (TL) monolingual corpora, and still enabling a similar functionality to that offered by conventional TM-based CAT tools. Our approach relies on cross-lingual sentence embeddings to retrieve translation proposals from TL monolingual corpora, and on a neural model to estimate their post-editing effort. The paper presents an automatic evaluation of these techniques on four language pairs that shows that our approach can successfully exploit monolingual texts in a TM-based CAT environment, increasing the amount of useful translation proposals, and that our neural model for estimating the post-editing effort enables the combination of translation proposals obtained from monolingual corpora and from TMs in the usual way. A human evaluation performed on a single language pair confirms the results of the automatic evaluation and seems to indicate that the translation proposals retrieved with our approach are more useful than what the automatic evaluation shows.

Pre-trained models have revolutionized the natural language processing field by leveraging large-scale language representations for various tasks. Some pre-trained models offer general-purpose representations, while others are specialized in particular tasks, like neural machine translation (NMT). Multilingual NMT-targeted systems are often fine-tuned for specific language pairs, but there is a lack of evidence-based best-practice recommendations to guide this process. Additionally, deploying these large pre-trained models in computationally restricted environments, typically found in developing regions where low-resource languages are spoken, has become challenging. We propose a pipeline to tune the mBART50 pre-trained model to 8 diverse low-resource language pairs, and then distill the resulting system to obtain lightweight and more sustainable NMT models. Our pipeline conveniently exploits back-translation, synthetic corpus filtering, and knowledge distillation to deliver efficient bilingual translation models that are 13 times smaller, while maintaining a close BLEU performance.

Los sistemas de juez en línea son aplicaciones web que se utilizan para la evaluación automática de algoritmos. Permiten a los usuarios enviar sus soluciones a distintos desafíos algorítmicos y se encargan de comprobar su validez, exigiendo a su vez que el tiempo de ejecución no supere un límite establecido. Estos sistemas desempeñan un papel fundamental en el ámbito de la educación y los concursos de programación. Este trabajo se centra en el desarrollo de un sistema de juez en línea, con el lenguaje de programación Python, que sea capaz de verificar la solución introducida y estimar su complejidad algorítmica asintótica en el caso promedio. Esta se obtendrá de forma empírica, evaluando la función del usuario con tamaños de entrada cada vez mayores. Al no contar con un límite de tiempo tan estricto y mostrar la complejidad, el sistema se convertirá en un recurso más didáctico que competitivo. Analizando diferentes algoritmos, hemos comprobado que la estimación del orden de complejidad realizada por el juez en línea coincide, en la mayoría de los casos, con el orden teórico. Por otro lado, este análisis también nos ha servido para comparar diferentes soluciones de un mismo desafío y confirmar que la mejor es la ofrecida por Python. Hemos validado la eficiencia de determinados operadores y funciones de Python, como in y sort, que permiten, respectivamente, buscar un elemento en una lista y ordenarla.

When the amount of parallel sentences available to train a neural machine translation is scarce, a common practice is to generate new synthetic training samples from them. A number of approaches have been proposed to produce synthetic parallel sentences that are similar to those in the parallel data available. These approaches work under the assumption that non-fluent target-side synthetic training samples can be harmful and may deteriorate translation performance. Even so, in this paper we demonstrate that synthetic training samples with non-fluent target sentences can improve translation performance if they are used in a multilingual machine translation framework as if they were sentences in another language. We conducted experiments on ten low-resource and four high-resource translation tasks and found out that this simple approach consistently improves translation performance as compared to state-of-the-art methods for generating synthetic training samples similar to those found in corpora. Furthermore, this improvement is independent of the size of the original training corpus, the resulting systems are much more robust against domain shift and produce less hallucinations.

Online Judge (OJ) systems are typically considered within programming-related courses as they yield fast and objective assessments of the code developed by the students. Such an evaluation generally provides a single decision based on a rubric, most commonly whether the submission successfully accomplished the assignment. Nevertheless, since in an educational context such information may be deemed insufficient, it would be beneficial for both the student and the instructor to receive additional feedback about the overall development of the task. This work aims to tackle this limitation by considering the further exploitation of the information gathered by the OJ and automatically inferring feedback for both the student and the instructor. More precisely, we consider the use of learning-based schemes—particularly, Multi-Instance Learning and classical Machine Learning formulations—to model student behaviour. Besides, Explainable Artificial Intelligence is contemplated to provide human-understandable feedback. The proposal has been evaluated considering a case of study comprising 2,500 submissions from roughly 90 different students from a programming-related course in a Computer Science degree. The results obtained validate the proposal: the model is capable of significantly predicting the user outcome (either passing or failing the assignment) solely based on the behavioural pattern inferred by the submissions provided to the OJ. Moreover, the proposal is able to identify prone-to-fail student groups and profiles as well as other relevant information, which eventually serves as feedback to both the student and the instructor.

Computer-aided translation (CAT) tools based on translation memories (MT) play a prominent role in the translation workflow of professional translators. However, the reduced availability of in-domain TMs, as compared to in-domain monolingual corpora, limits its adoption for a number of translation tasks. In this paper, we introduce a novel neural approach aimed at overcoming this limitation by exploiting not only TMs, but also in-domain target-language (TL) monolingual corpora, and still enabling a similar functionality to that offered by conventional TM-based CAT tools. Our approach relies on cross-lingual sentence embeddings to retrieve translation proposals from TL monolingual corpora, and on a neural model to estimate their post-editing effort. The paper presents an automatic evaluation of these techniques on four language pairs that shows that our approach can successfully exploit monolingual texts in a TM-based CAT environment, increasing the amount of useful translation proposals, and that our neural model for estimating the post-editing effort enables the combination of translation proposals obtained from monolingual corpora and from TMs in the usual way. A human evaluation performed on a single language pair confirms the results of the automatic evaluation and seems to indicate that the translation proposals retrieved with our approach are more useful than what the automatic evaluation shows.

Los sistemas de juez en línea (JL) se utilizan en el ámbito de la programación, ya que permiten realizar evaluaciones rápidas y objetivas del código desarrollado por los estudiantes. Este tipo de evaluación suele proporcionar una única decisión sobre si el envío ha cumplido con éxito la tarea. Sin embargo, dado que en un contexto educativo dicha información puede considerarse insuficiente, sería beneficioso tanto para el estudiante como para el instructor recibir información adicional sobre el desarrollo general de la tarea. Este trabajo pretende abordar esta limitación considerando la explotación adicional de la información recopilada por el JL e infiriendo automáticamente la retroalimentación, tanto para el estudiante como para el instructor. Consideramos el uso de esquemas basados en Aprendizaje Automático para modelar el comportamiento del estudiante y de Inteligencia Artificial Explicable para proporcionar una retroalimentación comprensible para el ser humano. La propuesta ha sido evaluada en un caso de estudio considerando más de 2.600 envíos y 90 estudiantes de la asignatura Desafíos de Programación (Ingeniería en Informática, Universidad de Alicante). Los resultados obtenidos validan la propuesta: el modelo es capaz de predecir de forma significativa el resultado del usuario (aprobar o suspender la tarea) basándose únicamente en el patrón de comportamiento inferido por los envíos realizados al JL, además retroalimentar al estudiante y al instructor.

In the context of neural machine translation, data augmentation (DA) techniques may be used for generating additional training samples when the available parallel data are scarce. Many DA approaches aim at expanding the support of the empirical data distribution by generating new sentence pairs that contain infrequent words, thus making it closer to the true data distribution of parallel sentences. In this paper, we propose to follow a completely different approach and present a multi-task DA approach in which we generate new sentence pairs with transformations, such as reversing the order of the target sentence, which produce unfluent target sentences. During training, these augmented sentences are used as auxiliary tasks in a multi-task framework with the aim of providing new contexts where the target prefix is not informative enough to predict the next word. This strengthens the encoder and forces the decoder to pay more attention to the source representations of the encoder. Experiments carried out on six low-resource translation tasks show consistent improvements over the baseline and over DA methods aiming at extending the support of the empirical data distribution. The systems trained with our approach rely more on the source tokens, are more robust against domain shift and suffer less hallucinations.

This paper studies the effects of word-level linguistic annotations in under-resourced neural machine translation, for which there is incomplete evidence in the literature. The study covers eight language pairs, different training corpus sizes, two architectures and three types of annotation: dummy tags (with no linguistic information at all), part-of-speech tags, and morpho-syntactic description tags, which consist of part of speech and morphological features. These linguistic annotations are interleaved in the input or output streams as a single tag placed before each word. In order to measure the performance under each scenario, we use automatic evaluation metrics and perform automatic error classification. Our experiments show that, in general, source-language annotations are helpful and morpho-syntactic descriptions outperform part of speech for some language pairs. On the contrary, when words are annotated in the target language, part-of-speech tags systematically outperform morpho-syntactic description tags in terms of automatic evaluation metrics, even though the use of morpho-syntactic description tags improves the grammaticality of the output. We provide a detailed analysis of the reasons behind this result.

Gravació de la sessió de pràctiques L16 de l'asignatura Tecnologies de la Traducció. Explicació de com utilitzar l'eina OmegaT en un projecte de traducció col·laborativa executat per l'alumnat.

Gravació de la sessió de pràctiques L15 de l'asignatura Tecnologies de la Traducció. Explicació de l'estratègia per triar la millor tecnologia/combinació de tecnologies en un projecte de traducció. Aquesta estimació forma part del desenvolupament d'un projecte de traducció col·laborativa executat per l'alumnat.

Gravació de la sessió pràctica L13. Explicació sobre cum dur a terme l'estimació de cost d'un projecte de traducció que utilitza tecnologies de traducció automàtica. Aquesta sessió pràctica forma part del projecte col·laboratiu de traducció que l'alumnat de l'assignatura ha d'executar.

Gravació de la sessió de pràctiques L18 de l'asignatura Tecnologies de la Traducció. Explicació de com dur a terme el control de qualitat en un projecte de traducció col·laborativa executat per l'alumnat.

This paper describes the process we followed in order to build English-Greek machine translation systems for the tourism/culture domain. We experimented with different data sets and domain adaptation methods for statistical machine translation and also built neural machine translation systems. The in-domain data were obtained by means of the ILSP Focused Crawler.

Gravació de la sessió de pràctiques L14 de l'asignatura Tecnologies de la Traducció. Explicació de l'estimació de cost d'un projecte de traducció que utilitza tecnologies de memòries de traducció. Aquesta estimació forma part del desenvolupament d'un projecte de traducció col·laborativa executat per l'alumnat.

Árboles AVL. Explicación detallada del algoritmo de borrado.

Ejercicios resueltos sobre inserciones y borrados en colas de prioridad implementadas mediante la estructura de datos montículo.

Árboles AVL. Ejercicios resueltos sobre borrado de elementos. Descripción detallada de los pasos a seguir y rotaciones a realizar.

Árboles AVL. Ejercicios resueltos de inserción de elementos, con explicación paso a paso de las rotaciones a realizar.

Definción del tipo abstracto de datos cola de prioridad. Explicación de las diferentes alternativas para su implementación y sus complejidades temporales. Definición de la implementación mediante montículo o HEAP.

Implementación de montículos mediante vectores. Algoritmo de ordenación Heapsort.

Introducción a los árboles AVL. Algoritmo de inserción.

Guía que pretende facilitar al alumnado de la asignatura Programación y Estrucuras de datos la implementación de un árbol equilibrado con respecto a la altura o AVL.

This paper presents a quantitative fine-grained manual evaluation approach to comparing the performance of different machine translation (MT) systems. We build upon the well-established multidimensional quality metrics (MQM) error taxonomy and implement a novel method that assesses whether the differences in performance for MQM error types between different MT systems are statistically significant. We conduct a case study for English-to-Croatian, a language direction that involves translating into a morphologically rich language, for which we compare three MT systems belonging to different paradigms: pure phrase-based, factored phrase-based and neural. First, we design an MQM-compliant error taxonomy tailored to the relevant linguistic phenomena of Slavic languages, which made the annotation process feasible and accurate. Errors in MT outputs were then annotated by two annotators following this taxonomy. Subsequently, we carried out a statistical analysis which showed that the best-performing system (neural) reduces the errors produced by the worst system (pure phrase-based) by more than half (54%). Moreover, we conducted an additional analysis of agreement errors in which we distinguished between short (phrase-level) and long distance (sentence-level) errors. We discovered that phrase-based MT approaches are of limited use for long distance agreement phenomena, for which neural MT was found to be especially effective.

This paper presents a new method with which to assist individuals with no background in linguistics to create monolingual dictionaries such as those used by the morphological analysers of many natural language processing applications. The involvement of non-expert users is especially critical for under-resourced languages which either lack or cannot afford the recruitment of a skilled workforce. Adding a word to a morphological dictionary usually requires identifying its stem along with the inflection paradigm that can be used in order to generate all the word forms of the new entry. Our method works under the assumption that the average speakers of a language can successfully answer the polar question “is x a valid form of the word w to be inserted?”, where x represents tentative alternative (inflected) forms of the new word w. The experiments show that with a small number of polar questions the correct stem and paradigm can be obtained from non-experts with high success rates. We study the impact of different heuristic and probabilistic approaches on the actual number of questions.