Search

International audience; Translation relations, which distinguish literal translation from other translation techniques, constitute an important subject of study for human translators (Chuquet and Paillard, 1989). However , automatic processing techniques based on interlingual relations, such as machine translation or paraphrase generation exploiting translational equivalence, have not made use of these relations explicitly until now. In this work, we present a categorization of translation relations and then we annotate a parallel multilingual (English, French, Chinese) corpus of oral presentations, the TED Talks, with these relations. Our long-term objective will be to automatically detect these relations in order to integrate them as important characteristics for the search of monolingual segments in relation of equivalence (paraphrases) or of entailment. The annotated corpus resulting from our work will be made available to the community.

International audience; This paper proposes a fine-grained flexible analysis methodology to reveal the residual difficulties of a high-quality Statistical Machine Translation (SMT) system. This proposal is motivated by the fact that the traditional automated metrics are not enough informative to indicate the nature and reasons of those residual difficulties. Their resolution is however a key point towards improving the high-quality output. The novelty of our approach consists in diagnosing Machine Translation (MT) performance by making a connection between errors, the characteristics of source sentences and some internal parameters of the system, using traces of Post-Edition (PE) operations as well as Quality Estimation (QE) techniques. Our methodology is illustrated on a SMT system adapted to the medical domain, based on a high quality English-French parallel corpus of Cochrane systematic review abstracts. Our experimental results show that the main difficulties that the system faces are in the domains of term precision and source language syntactic and stylistic peculiarities. We furthermore provide general information regarding the corpus structure and its specificities, including internal stylistic varieties characteristic of this sub-genre.

International audience; Literary works are becoming increasingly available in electronic for- mats, thus quickly transforming editorial processes and reading habits. In the context of the global enthusiasm for multilingualism, the rapid spread of e-book readers, such as Amazon Kindle or Kobo Touch , fosters the development of a new generation of reading tools for biingual books. In particular, literary works, when available in several languages, offer an attractive perspective for self-development or everyday leisure reading, but also for activities such as language learning, translation or literary studies.An important issue in the automatic processing of multilingual e- books is the alignment between textual units. Alignment could help identify corresponding text units in different languages, which would be particularly beneficial to bilingual readers and translation profes- sionals. Computing automatic alignments for literary works, however, is a task more challenging than in the case of better behaved corpora such as parliamentary proceedings or technical manuals. In this paper, we revisit the problem of computing high-quality alignment for literary works. We first perform a large-scale evaluation of automatic alignment for literary texts, which provides a fair assessment of the actual difficulty of this task. We then introduce a two-pass approach, based on a maximum entropy model. Experimental results for novels available in English and French or in English and Spanish demonstrate the effectiveness of our method.

Dans cet article de démonstration, nous introduisons un logiciel permettant de construire des tables de traduction de manière beaucoup plus rapide que ne le font les techniques de l'état de l'art Cette accélération notable est obtenue par le biais d'un double échantillonnage : l'un permet la sélection d'un nombre limité de bi-phrases contenant les segments à traduire, l'autre réalise un alignement à la volée de ces bi-phrases pour extraire des exemples de traduction.

Dans cet article, nous montrons comment l'utilisation conjointe d'une technique d'alignement de phrases parallèles à la demande et d'estimation de modèles de traduction à la volée permet une réduction en temps très notable (jusqu'à 93% dans nos expériences) par rapport à un système à l'état de l'art, tout en offrant un compromis en termes de qualité très intéressant dans certaines configurations. En particulier, l'exploitation immédiate de documents traduits permet de compenser très rapidement l'absence d'un corpus de développement.

International audience; abstract

International audience; This paper describes LIMSI's submissions to the shared WMT'13 translation task. We report results for French-English, German-English and Spanish-English in both directions. Our submissions use n-code, an open source system based on bilingual n-grams, and continuous space models in a post-processing step. The main novelties of this year's participation are the following: our first participation to the Spanish-English task; experiments with source pre-ordering; a tighter integration of continuous space language models using artificial text generation (for Ger-man); and the use of different tuning sets according to the original language of the text to be translated.

International audience; Electronic versions of literary works abound on the Internet and the rapid dissemination of electronic readers will make electronic books more and more common. It is often the case thatliterary works exist in more than one language, suggesting that, if properly aligned, they could be turned into useful resources for many practical applications, such as writing and language learning aids, translation studies, or data-based machine translation. To be of any use, these bilingual works need to be aligned as precisely as possible, a notoriously difficult task. In this paper, we revisit the problem of sentence alignment for literary works and explore the performance of a new, multi-pass, approach based on a combination of systems. Experiments conducted on excerpts of ten masterpieces of the French and English literature show that our approach significantly outperforms two open source tools.

International audience; abstract

Ouvrage consultable à cette adresse : http://www.statmt.org/wmt11/WMT-2011.pdf; International audience; This paper describes LIMSI's submissions to the Sixth Workshop on Statistical Machine Translation. We report results for the French-English and German-English shared translation tasks in both directions. Our systems use n-code, an open source Statistical Machine Translation system based on bilingual n-grams. For the French-English task, we focussed on finding efficient ways to take advantage of the large and heterogeneous training parallel data. In particular, using a simple filtering strategy helped to improve both processing time and translation quality. To translate from English to French and German, we also investigated the use of the SOUL language model in Machine Translation and showed significant improvements with a 10-gram SOUL model. We also briefly report experiments with several alternatives to the standard n-best MERT procedure, leading to a significant speed-up.

International audience

En éste artículo, se analizan las modificaciones accesibles a través del historial de revisiones de Wikipedia en francés. Se define una tipología de modificaciones basada en el estudio detallado de WiCoPaCo, un recurso gratuito construido a través de un estudio del historial de revisiones de Wikipedia. Conforme a esta tipología, detallamos el estudio de la anotación manual de un subconjunto del corpus, con la intención de evaluar la dificultad de la tarea de identificación automática de paráfrasis en el mismo corpus. Finalmente, evaluamos una herramienta de identificación de paráfrasis a base de reglas. In this article, we analyse the modifications available in the French Wikipedia revision history. We define a typology of modifications based on a detailed study of WiCoPaCo, a freely-available resource built by automatically mining Wikipedia’s revision history. Based on this typology, we detail a manual annotation study of a subpart of the corpus aimed at assessing the difficulty of automatic paraphrase identification in such a corpus. Finally, we assess a rule-based paraphrase identification tool. This work was supported by a grant from LIMSI.

Rephrasing text spans is a common task when revising a text. However, traditional dictionaries often cannot provide direct assistance to writers in performing this task. In this article, we describe an approach to obtain a monolingual phrase lexicon using techniques used in Statistical Machine Translation. A part to be rephrased is first translated into a pivot language, and then translated back into the original language. Models for assessing fluency, meaning preservation and lexical divergence are used to rank possible rephrasings, and their relative weight can be tuned by the user so as to better address her needs. An evaluation shows that these models can be used successfully to select rephrasings that are likely to be useful to a writer.

@inproceedings{CI-HERNANDEZ-2006-2, author = {Nicolas Hernandez and Aur{é}lien Max and Michael Zock}, title = {{Support for interactive text structuring in word processors}}, booktitle = {Proceedings of the Writing Development in Higher Education (WDHE) Conference}, year = {2006}, address = {Open University, Milton Keynes, UK}, month = {May} }; International audience