Emmanuelle Esperança-Rodier

Also published as: Emmanuelle Esperanca-Rodier


2024

pdf bib
Exploring NMT Explainability for Translators Using NMT Visualising Tools
Gabriela Gonzalez-Saez | Mariam Nakhle | James Turner | Fabien Lopez | Nicolas Ballier | Marco Dinarelli | Emmanuelle Esperança-Rodier | Sui He | Raheel Qader | Caroline Rossi | Didier Schwab | Jun Yang
Proceedings of the 25th Annual Conference of the European Association for Machine Translation (Volume 1)

This paper describes work in progress on Visualisation tools to foster collaborations between translators and computational scientists. We aim to describe how visualisation features can be used to explain translation and NMT outputs. We tested several visualisation functionalities with three NMT models based on Chinese-English, Spanish-English and French-English language pairs. We created three demos containing different visualisation tools and analysed them within the framework of performance-explainability, focusing on the translator’s perspective.

pdf bib
The MAKE-NMTViz Project: Meaningful, Accurate and Knowledge-limited Explanations of NMT Systems for Translators
Gabriela Gonzalez-Saez | Fabien Lopez | Mariam Nakhle | James Turner | Nicolas Ballier | Marco Dinarelli | Emmanuelle Esperança-Rodier | Sui He | Caroline Rossi | Didier Schwab | Jun Yang
Proceedings of the 25th Annual Conference of the European Association for Machine Translation (Volume 2)

This paper describes MAKE-NMTViz, a project designed to help translators visualize neural machine translation outputs using explainable artificial intelligence visualization tools initially developed for computer vision.

pdf bib
Approches cascade et de bout-en-bout pour la traduction automatique de la parole en pictogrammes
Cécile Macaire | Chloé Dion | Didier Schwab | Benjamin Lecouteux | Emmanuelle Esperança-Rodier
Actes de la 31ème Conférence sur le Traitement Automatique des Langues Naturelles, volume 1 : articles longs et prises de position

La traduction automatique de la parole en pictogrammes (Parole-à-Pictos) est une nouvelle tâche du Traitement Automatique des Langues (TAL) ayant pour but de proposer une séquence de pictogrammes à partir d’un énoncé oral. Cet article explore deux approches distinctes : (1) en cascade, qui combine un système de reconnaissance vocale avec un système de traduction, et (2) de bout-en-bout, qui adapte un système de traduction automatique de la parole. Nous comparons différentes architectures état de l’art entraînées sur nos propres données alignées parole-pictogrammes. Nous présentons une première évaluation automatique des systèmes et réalisons une évaluation humaine pour analyser leur comportement et leur impact sur la traduction en pictogrammes. Les résultats obtenus mettent en évidence la capacité d’une approche en cascade à générer des traductions acceptables à partir de la parole lue et dans des contextes de la vie quotidienne.

pdf bib
Sur les limites de l’identification par l’humain de textes générés automatiquement
Nadége Alavoine | Maximin Coavoux | Emmanuelle Esperança-Rodier | Romane Gallienne | Carlos-Emiliano González-Gallardo | Jérôme Goulian | Jose G Moreno | Aurélie Névéol | Didier Schwab | Vincent Segonne | Johanna Simoens
Actes de la 31ème Conférence sur le Traitement Automatique des Langues Naturelles, volume 2 : traductions d'articles publiès

La génération de textes neuronaux fait l’objet d’une grande attention avec la publication de nouveaux outils tels que ChatGPT. La principale raison en est que la qualité du texte généré automatiquement peut être attribuée à un$cdot$e rédacteurice humain$cdot$e même quand l’évaluation est faite par un humain. Dans cet article, nous proposons un nouveau corpus en français et en anglais pour la tâche d’identification de textes générés automatiquement et nous menons une étude sur la façon dont les humains perçoivent ce texte. Nos résultats montrent, comme les travaux antérieurs à l’ère de ChatGPT, que les textes générés par des outils tels que ChatGPT partagent certaines caractéristiques communes mais qu’ils ne sont pas clairement identifiables, ce qui génère des perceptions différentes de ces textes par l’humain.

pdf bib
Un corpus multimodal alignant parole, transcription et séquences de pictogrammes dédié à la traduction automatique de la parole vers des pictogrammes
Cécile Macaire | Chloé Dion | Jordan Arrigo | Claire Lemaire | Emmanuelle Esperança-Rodier | Benjamin Lecouteux | Didier Schwab
Actes de la 31ème Conférence sur le Traitement Automatique des Langues Naturelles, volume 2 : traductions d'articles publiès

La traduction automatique de la parole vers des pictogrammes peut faciliter la communication entre des soignants et des personnes souffrant de troubles du langage. Cependant, il n’existe pas de formalisme de traduction établi, ni d’ensembles de données accessibles au public pour l’entraînement de systèmes de traduction de la parole vers des pictogrammes. Cet article présente le premier ensemble de données alignant de la parole, du texte et des pictogrammes. Ce corpus comprend plus de 230 heures de parole. Nous discutons de nos choix pour créer une grammaire adaptée à des séquences de pictogrammes. Cette dernière s’articule autour de règles et d’un vocabulaire restreint. La grammaire résulte d’une étude linguistique approfondie des ressources extraites du site Web d’ARASAAC. Nous avons ensuite validé ces règles à l’issue de multiples phases de post-édition par des annotateurs experts. Le corpus proposé est ensuite utilisé pour entraîner un système en cascade traduisant la parole vers des pictogrammes. L’ensemble du corpus est disponible gratuitement sur le site web d’Ortolang sous une licence non commerciale. Il s’agit d’un point de départ pour la recherche portant sur la traduction automatique de la parole vers des pictogrammes.

pdf bib
A Multimodal French Corpus of Aligned Speech, Text, and Pictogram Sequences for Speech-to-Pictogram Machine Translation
Cécile Macaire | Chloé Dion | Jordan Arrigo | Claire Lemaire | Emmanuelle Esperança-Rodier | Benjamin Lecouteux | Didier Schwab
Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)

The automatic translation of spoken language into pictogram units can facilitate communication involving individuals with language impairments. However, there is no established translation formalism or publicly available datasets for training end-to-end speech translation systems. This paper introduces the first aligned speech, text, and pictogram translation dataset ever created in any language. We provide a French dataset that contains 230 hours of speech resources. We create a rule-based pictogram grammar with a restricted vocabulary and include a discussion of the strategic decisions involved. It takes advantage of an in-depth linguistic study of resources taken from the ARASAAC website. We validate these rules through multiple post-editing phases by expert annotators. The constructed dataset is then used to experiment with a Speech-to-Pictogram cascade model, which employs state-of-the-art Automatic Speech Recognition models. The dataset is freely available under a non-commercial licence. This marks a starting point to conduct research into the automatic translation of speech into pictogram units.

pdf bib
Jargon: A Suite of Language Models and Evaluation Tasks for French Specialized Domains
Vincent Segonne | Aidan Mannion | Laura Cristina Alonzo Canul | Alexandre Daniel Audibert | Xingyu Liu | Cécile Macaire | Adrien Pupier | Yongxin Zhou | Mathilde Aguiar | Felix E. Herron | Magali Norré | Massih R Amini | Pierrette Bouillon | Iris Eshkol-Taravella | Emmanuelle Esperança-Rodier | Thomas François | Lorraine Goeuriot | Jérôme Goulian | Mathieu Lafourcade | Benjamin Lecouteux | François Portet | Fabien Ringeval | Vincent Vandeghinste | Maximin Coavoux | Marco Dinarelli | Didier Schwab
Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)

Pretrained Language Models (PLMs) are the de facto backbone of most state-of-the-art NLP systems. In this paper, we introduce a family of domain-specific pretrained PLMs for French, focusing on three important domains: transcribed speech, medicine, and law. We use a transformer architecture based on efficient methods (LinFormer) to maximise their utility, since these domains often involve processing long documents. We evaluate and compare our models to state-of-the-art models on a diverse set of tasks and datasets, some of which are introduced in this paper. We gather the datasets into a new French-language evaluation benchmark for these three domains. We also compare various training configurations: continued pretraining, pretraining from scratch, as well as single- and multi-domain pretraining. Extensive domain-specific experiments show that it is possible to attain competitive downstream performance even when pre-training with the approximative LinFormer attention mechanism. For full reproducibility, we release the models and pretraining data, as well as contributed datasets.

pdf bib
Limitations of Human Identification of Automatically Generated Text
Nadège Alavoine | Maximin Coavoux | Emmanuelle Esperança-Rodier | Romane Gallienne | Carlos-Emiliano González-Gallardo | Jérôme Goulian | Jose G. Moreno | Aurélie Névéol | Didier Schwab | Vincent Segonne | Johanna Simoens
Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)

Neural text generation is receiving broad attention with the publication of new tools such as ChatGPT. The main reason for that is that the achieved quality of the generated text may be attributed to a human writer by the naked eye of a human evaluator. In this paper, we propose a new corpus in French and English for the task of recognising automatically generated texts and we conduct a study of how humans perceive the text. Our results show, as previous work before the ChatGPT era, that the generated texts by tools such as ChatGPT share some common characteristics but they are not clearly identifiable which generates different perceptions of these texts.

2023

pdf bib
Plateformes pour la création de données en pictogrammes
Cécile Macaire | Jordan Arrigo | Chloé Dion | Claire Lemaire | Emmanuelle Esperança-Rodier | Benjamin Lecouteux | Didier Schwab
Actes de CORIA-TALN 2023. Actes de la 30e Conférence sur le Traitement Automatique des Langues Naturelles (TALN), volume 5 : démonstrations

Nous présentons un ensemble de trois interfaces web pour la création de données en pictogrammes dans le cadre du projet ANR Propicto. Chacune a un objectif précis : annoter des données textuelles en pictogrammes ARASAAC, créer un vocabulaire en pictogrammes, et post-éditer des phrases annotées en pictogrammes. Bien que nécessaire pour des outils de traduction automatique vers les unités pictographiques, actuellement, presque aucune ressource annotée n’existe. Cet article présente les spécificités de ces plateformes web (disponibles en ligne gratuitement) et leur utilité.

pdf bib
Voice2Picto : un système de traduction automatique de la parole vers des pictogrammes
Cécile Macaire | Emmanuelle Esperança-Rodier | Benjamin Lecouteux | Didier Schwab
Actes de CORIA-TALN 2023. Actes de la 30e Conférence sur le Traitement Automatique des Langues Naturelles (TALN), volume 5 : démonstrations

Nous présentons Voice2Picto, un système de traduction permettant, à partir de l’oral, de proposer une séquence de pictogrammes correspondants. S’appuyant sur des technologies du traitement automatique du langage naturel, l’outil a deux objectifs : améliorer l’accès à la communication pour (1) les personnes allophones dans un contexte d’urgence médicale, et (2) pour les personnes avec des difficultés de parole. Il permettra aux personnes des services hospitaliers, et aux familles de véhiculer un message en pictogrammes facilement compréhensible auprès de personnes ne pouvant communiquer via les canaux traditionnels de communication (parole, gestes, langue des signes). Dans cet article, nous décrivons l’architecture du système de Voice2Picto et les pistes futures. L’application est en open-source via un dépôt Git : https://github.com/macairececile/Voice2Picto.

pdf bib
PROPICTO: Developing Speech-to-Pictograph Translation Systems to Enhance Communication Accessibility
Lucía Ormaechea | Pierrette Bouillon | Maximin Coavoux | Emmanuelle Esperança-Rodier | Johanna Gerlach | Jerôme Goulian | Benjamin Lecouteux | Cécile Macaire | Jonathan Mutal | Magali Norré | Adrien Pupier | Didier Schwab
Proceedings of the 24th Annual Conference of the European Association for Machine Translation

PROPICTO is a project funded by the French National Research Agency and the Swiss National Science Foundation, that aims at creating Speech-to-Pictograph translation systems, with a special focus on French as an input language. By developing such technologies, we intend to enhance communication access for non-French speaking patients and people with cognitive impairments.

pdf bib
The MAKE-NMTVIZ System Description for the WMT23 Literary Task
Fabien Lopez | Gabriela González | Damien Hansen | Mariam Nakhle | Behnoosh Namdarzadeh | Nicolas Ballier | Marco Dinarelli | Emmanuelle Esperança-Rodier | Sui He | Sadaf Mohseni | Caroline Rossi | Didier Schwab | Jun Yang | Jean-Baptiste Yunès | Lichao Zhu
Proceedings of the Eighth Conference on Machine Translation

This paper describes the MAKE-NMTVIZ Systems trained for the WMT 2023 Literary task. As a primary submission, we used Train, Valid1, test1 as part of the GuoFeng corpus (Wang et al., 2023) to fine-tune the mBART50 model with Chinese-English data. We followed very similar training parameters to (Lee et al. 2022) when fine-tuning mBART50. We trained for 3 epochs, using gelu as an activation function, with a learning rate of 0.05, dropout of 0.1 and a batch size of 16. We decoded using a beam search of size 5. For our contrastive1 submission, we implemented a fine-tuned concatenation transformer (Lupo et al., 2023). The training was developed in two steps: (i) a sentence-level transformer was implemented for 10 epochs trained using general, test1, and valid1 data (more details in contrastive2 system); (ii) second, we fine-tuned at document-level using 3-sentence concatenation for 4 epochs using train, test2, and valid2 data. During the fine-tuning, we used ReLU as an activation function, with an inverse square root learning rate, dropout of 0.1, and a batch size of 64. We decoded using a beam search of size. Four our contrastive2 and last submission, we implemented a sentence-level transformer model (Vaswani et al., 2017). The model was trained with general data for 10 epochs using general-purpose, test1, and valid 1 data. The training parameters were an inverse square root scheduled learning rate, a dropout of 0.1, and a batch size of 64. We decoded using a beam search of size 4. We then compared the three translation outputs from an interdisciplinary perspective, investigating some of the effects of sentence- vs document-based training. Computer scientists, translators and corpus linguists discussed the linguistic remaining issues for this discourse-level literary translation.

2021

pdf bib
ACCOLÉ : Annotation Collaborative d’erreurs de traduction pour COrpus aLignÉs, multi-cibles, et Annotation d’Expressions Poly-lexicales (ACCOLÉ: A Collaborative Platform of Error Annotation for Aligned)
Emmanuelle Esperança-Rodier | Francis Brunet-Manquat
Actes de la 28e Conférence sur le Traitement Automatique des Langues Naturelles. Volume 3 : Démonstrations

Cette démonstration présente les avancées d’ACCOLÉ (Annotation Collaborative d’erreurs de traduction pour COrpus aLignÉs), qui en plus de proposer une gestion simplifiée des corpus et des typologies d’erreurs, l’annotation d’erreurs pour des corpus de traduction bilingues alignés, la collaboration et/ou supervision lors de l’annotation, la recherche de modèle d’erreurs dans les annotations, permet désormais d’annoter les Expressions Polylexicales (EPL) dans des textes monolingues en français, et d’accéder à l’annotation d’erreurs pour des corpus de traduction multicibles. Dans cet article, après un bref rappel des fonctionnalités d’ACCOLÉ, nous explicitons les fonctionnalités de chaque nouveauté.

2020

pdf bib
Online Versus Offline NMT Quality: An In-depth Analysis on English-German and German-English
Maha Elbayad | Michael Ustaszewski | Emmanuelle Esperança-Rodier | Francis Brunet-Manquat | Jakob Verbeek | Laurent Besacier
Proceedings of the 28th International Conference on Computational Linguistics

We conduct in this work an evaluation study comparing offline and online neural machine translation architectures. Two sequence-to-sequence models: convolutional Pervasive Attention (Elbayad et al. 2018) and attention-based Transformer (Vaswani et al. 2017) are considered. We investigate, for both architectures, the impact of online decoding constraints on the translation quality through a carefully designed human evaluation on English-German and German-English language pairs, the latter being particularly sensitive to latency constraints. The evaluation results allow us to identify the strengths and shortcomings of each model when we shift to the online setup.

pdf bib
Providing Semantic Knowledge to a Set of Pictograms for People with Disabilities: a Set of Links between WordNet and Arasaac: Arasaac-WN
Didier Schwab | Pauline Trial | Céline Vaschalde | Loïc Vial | Emmanuelle Esperanca-Rodier | Benjamin Lecouteux
Proceedings of the Twelfth Language Resources and Evaluation Conference

This article presents a resource that links WordNet, the widely known lexical and semantic database, and Arasaac, the largest freely available database of pictograms. Pictograms are a tool that is more and more used by people with cognitive or communication disabilities. However, they are mainly used manually via workbooks, whereas caregivers and families would like to use more automated tools (use speech to generate pictograms, for example). In order to make it possible to use pictograms automatically in NLP applications, we propose a database that links them to semantic knowledge. This resource is particularly interesting for the creation of applications that help people with cognitive disabilities, such as text-to-picto, speech-to-picto, picto-to-speech... In this article, we explain the needs for this database and the problems that have been identified. Currently, this resource combines approximately 800 pictograms with their corresponding WordNet synsets and it is accessible both through a digital collection and via an SQL database. Finally, we propose a method with associated tools to make our resource language-independent: this method was applied to create a first text-to-picto prototype for the French language. Our resource is distributed freely under a Creative Commons license at the following URL: https://github.com/getalp/Arasaac-WN.

2019

pdf bib
With or without post-editing processes? Evidence for a gap in machine translation evaluation
Caroline Rossi | Emmanuelle Esperança-Rodier
Proceedings of the Second MEMENTO workshop on Modelling Parameters of Cognitive Effort in Translation Production

2018

pdf bib
ACCOLÉ : Annotation Collaborative d’erreurs de traduction pour COrpus aLignÉs (ACCOLÉ: A Collaborative Platform of Error Annotation for Aligned Corpus)
Francis Brunet-Manquat | Emmanuelle Esperança-Rodier
Actes de la Conférence TALN. Volume 2 - Démonstrations, articles des Rencontres Jeunes Chercheurs, ateliers DeFT

La plateforme ACCOLÉ (Annotation Collaborative d’erreurs de traduction pour COrpus aLignÉs) propose une palette de services innovants permettant de répondre aux besoins modernes d’analyse d’erreurs de traduction : gestion simplifiée des corpus et des typologies d’erreurs, annotation d’erreurs efficace, collaboration et/ou supervision lors de l’annotation, recherche de modèle d’erreurs dans les annotations.

2016

bib
Translation quality evaluation of MWE from French into English using an SMT system
Emmanuelle Esperança-Rodier | Johan Didier
Proceedings of Translating and the Computer 38

2012

pdf bib
Collection of a Large Database of French-English SMT Output Corrections
Marion Potet | Emmanuelle Esperança-Rodier | Laurent Besacier | Hervé Blanchon
Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12)

Corpus-based approaches to machine translation (MT) rely on the availability of parallel corpora. To produce user-acceptable translation outputs, such systems need high quality data to be efficiency trained, optimized and evaluated. However, building high quality dataset is a relatively expensive task. In this paper, we describe the data collection and analysis of a large database of 10.881 SMT translation output hypotheses manually corrected. These post-editions were collected using Amazon's Mechanical Turk, following some ethical guidelines. A complete analysis of the collected data pointed out a high quality of the corrections with more than 87 % of the collected post-editions that improve hypotheses and more than 94 % of the crowdsourced post-editions which are at least of professional quality. We also post-edited 1,500 gold-standard reference translations (of bilingual parallel corpora generated by professional) and noticed that 72 % of these translations needed to be corrected during post-edition. We computed a proximity measure between the differents kind of translations and pointed out that reference translations are as far from the hypotheses than from the corrected hypotheses (i.e. the post-editions). In light of these last findings, we discuss the adequation of text-based generated reference translations to train setence-to-sentence based SMT systems.

2011

pdf bib
Oracle-based Training for Phrase-based Statistical Machine Translation
Marion Potet | Emmanuelle Esperança-Rodier | Hervé Blanchon | Laurent Besacier
Proceedings of the 15th Annual Conference of the European Association for Machine Translation