Guillaume Gravier

Also published as: G. Gravier


2021

pdf bib
Active Learning for Interactive Relation Extraction in a French Newspaper’s Articles
Cyrielle Mallart | Michel Le Nouy | Guillaume Gravier | Pascale Sébillot
Proceedings of the International Conference on Recent Advances in Natural Language Processing (RANLP 2021)

Relation extraction is a subtask of natural langage processing that has seen many improvements in recent years, with the advent of complex pre-trained architectures. Many of these state-of-the-art approaches are tested against benchmarks with labelled sentences containing tagged entities, and require important pre-training and fine-tuning on task-specific data. However, in a real use-case scenario such as in a newspaper company mostly dedicated to local information, relations are of varied, highly specific type, with virtually no annotated data for such relations, and many entities co-occur in a sentence without being related. We question the use of supervised state-of-the-art models in such a context, where resources such as time, computing power and human annotators are limited. To adapt to these constraints, we experiment with an active-learning based relation extraction pipeline, consisting of a binary LSTM-based lightweight model for detecting the relations that do exist, and a state-of-the-art model for relation classification. We compare several choices for classification models in this scenario, from basic word embedding averaging, to graph neural networks and Bert-based ones, as well as several active learning acquisition strategies, in order to find the most cost-efficient yet accurate approach in our French largest daily newspaper company’s use case.

2020

pdf bib
Relation, es-tu là ? Détection de relations par LSTM pour améliorer l’extraction de relations (Relation, are you there ? LSTM-based relation detection to improve knowledge extraction )
Cyrielle Mallart | Michel Le Nouy | Guillaume Gravier | Pascale Sébillot
Actes de la 6e conférence conjointe Journées d'Études sur la Parole (JEP, 33e édition), Traitement Automatique des Langues Naturelles (TALN, 27e édition), Rencontre des Étudiants Chercheurs en Informatique pour le Traitement Automatique des Langues (RÉCITAL, 22e édition). Volume 2 : Traitement Automatique des Langues Naturelles

De nombreuses méthodes d’extraction et de classification de relations ont été proposées et testées sur des données de référence. Cependant, dans des données réelles, le nombre de relations potentielles est énorme et les heuristiques souvent utilisées pour distinguer de vraies relations de co-occurrences fortuites ne détectent pas les signaux faibles pourtant importants. Dans cet article, nous étudions l’apport d’un modèle de détection de relations, identifiant si un couple d’entités dans une phrase exprime ou non une relation, en tant qu’étape préliminaire à la classification des relations. Notre modèle s’appuie sur le plus court chemin de dépendances entre deux entités, modélisé par un LSTM et combiné avec les types des entités. Sur la tâche de détection de relations, nous obtenons de meilleurs résultats qu’un modèle état de l’art pour la classification de relations, avec une robustesse accrue aux relations inédites. Nous montrons aussi qu’une détection binaire en amont d’un modèle de classification améliore significativement ce dernier.

pdf bib
On the Correlation of Word Embedding Evaluation Metrics
François Torregrossa | Vincent Claveau | Nihel Kooli | Guillaume Gravier | Robin Allesiardo
Proceedings of the 12th Language Resources and Evaluation Conference

Word embeddings intervene in a wide range of natural language processing tasks. These geometrical representations are easy to manipulate for automatic systems. Therefore, they quickly invaded all areas of language processing. While they surpass all predecessors, it is still not straightforward why and how they do so. In this article, we propose to investigate all kind of evaluation metrics on various datasets in order to discover how they correlate with each other. Those correlations lead to 1) a fast solution to select the best word embeddings among many others, 2) a new criterion that may improve the current state of static Euclidean word embeddings, and 3) a way to create a set of complementary datasets, i.e. each dataset quantifies a different aspect of word embeddings.

2017

pdf bib
Language-based Construction of Explorable News Graphs for Journalists
Rémi Bois | Guillaume Gravier | Eric Jamet | Emmanuel Morin | Pascale Sébillot | Maxime Robert
Proceedings of the 2017 EMNLP Workshop: Natural Language Processing meets Journalism

Faced with ever-growing news archives, media professionals are in need of advanced tools to explore the information surrounding specific events. This problem is most commonly answered by browsing news datasets, going from article to article and viewing unaltered original content. In this article, we introduce an efficient way to generate links between news items, allowing such browsing through an easily explorable graph, and enrich this graph by automatically typing links in order to inform the user on the nature of the relation between two news pieces. User evaluations are conducted on real world data with journalists in order to assess for the interest of both the graph representation and link typing in a press reviewing task, showing the system to be of significant help for their work.

2016

pdf bib
Évaluation dune nouvelle structuration thématique hiérarchique des textes dans un cadre de résumé automatique et de détection d’ancres au sein de vidéos (Evaluation of a novel hierarchical thematic structuring of texts in the framework of text summarization and anchor detection for video hyperlinking )
Anca Simon | Guillaume Gravier | Pascale Sébillot
Actes de la conférence conjointe JEP-TALN-RECITAL 2016. volume 2 : TALN (Articles longs)

automatique et de détection d’ancres au sein de vidéos Anca Simon1 Guillaume Gravier2 Pascale Sébillot3 (1) Université de Rennes 1, IRISA & INRIA Rennes, Campus de Beaulieu, 35042 Rennes, France (2) CNRS, IRISA & INRIA Rennes, Campus de Beaulieu, 35042 Rennes, France (3) INSA, IRISA & INRIA Rennes, Campus de Beaulieu, 35042 Rennes, France anca.simon@irisa.fr, guillaume.gravier@irisa.fr, pascale.sebillot@irisa.fr R ÉSUMÉ Dans cet article, nous évaluons, à travers son intérêt pour le résumé automatique et la détection d’ancres dans des vidéos, le potentiel d’une nouvelle structure thématique extraite de données textuelles, composée d’une hiérarchie de fragments thématiquement focalisés. Cette structure est produite par un algorithme exploitant les distributions temporelles d’apparition des mots dans les textes en se fondant sur une analyse de salves lexicales. La hiérarchie obtenue a pour objet de filtrer le contenu non crucial et de ne conserver que l’information saillante des textes, à différents niveaux de détail. Nous montrons qu’elle permet d’améliorer la production de résumés ou au moins de maintenir les résultats de l’état de l’art, tandis que pour la détection d’ancres, elle nous conduit à la meilleure précision dans le contexte de la tâche Search and Anchoring in Video Archives à MediaEval. Les expériences sont réalisées sur du texte écrit et sur un corpus de transcriptions automatiques d’émissions de télévision.

2015

pdf bib
Vers une typologie de liens entre contenus journalistiques
Remi Bois | Guillaume Gravier | Emmanuel Morin | Pascale Sébillot
Actes de la 22e conférence sur le Traitement Automatique des Langues Naturelles. Articles courts

Nous présentons une typologie de liens pour un corpus multimédia ancré dans le domaine journalistique. Bien que plusieurs typologies aient été créées et utilisées par la communauté, aucune ne permet de répondre aux enjeux de taille et de variété soulevés par l’utilisation d’un corpus large comprenant des textes, des vidéos, ou des émissions radiophoniques. Nous proposons donc une nouvelle typologie, première étape visant à la création et la catégorisation automatique de liens entre des fragments de documents afin de proposer de nouveaux modes de navigation au sein d’un grand corpus. Plusieurs exemples d’instanciation de la typologie sont présentés afin d’illustrer son intérêt.

pdf bib
Hierarchical Topic Structuring: From Dense Segmentation to Topically Focused Fragments via Burst Analysis
Anca-Roxana Simon | Pascale Sébillot | Guillaume Gravier
Proceedings of the International Conference Recent Advances in Natural Language Processing

2014

pdf bib
The ETAPE speech processing evaluation
Olivier Galibert | Jeremy Leixa | Gilles Adda | Khalid Choukri | Guillaume Gravier
Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC'14)

The ETAPE evaluation is the third evaluation in automatic speech recognition and associated technologies in a series which started with ESTER. This evaluation proposed some new challenges, by proposing TV and radio shows with prepared and spontaneous speech, annotation and evaluation of overlapping speech, a cross-show condition in speaker diarization, and new, complex but very informative named entities in the information extraction task. This paper presents the whole campaign, including the data annotated, the metrics used and the anonymized system results. All the data created in the evaluation, hopefully including system outputs, will be distributed through the ELRA catalogue in the future.

pdf bib
Bridging the gap between speech technology and natural language processing: an evaluation toolbox for term discovery systems
Bogdan Ludusan | Maarten Versteegh | Aren Jansen | Guillaume Gravier | Xuan-Nga Cao | Mark Johnson | Emmanuel Dupoux
Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC'14)

The unsupervised discovery of linguistic terms from either continuous phoneme transcriptions or from raw speech has seen an increasing interest in the past years both from a theoretical and a practical standpoint. Yet, there exists no common accepted evaluation method for the systems performing term discovery. Here, we propose such an evaluation toolbox, drawing ideas from both speech technology and natural language processing. We first transform the speech-based output into a symbolic representation and compute five types of evaluation metrics on this representation: the quality of acoustic matching, the quality of the clusters found, and the quality of the alignment with real words (type, token, and boundary scores). We tested our approach on two term discovery systems taking speech as input, and one using symbolic input. The latter was run using both the gold transcription and a transcription obtained from an automatic speech recognizer, in order to simulate the case when only imperfect symbolic information is available. The results obtained are analysed through the use of the proposed evaluation metrics and the implications of these metrics are discussed.

2013

pdf bib
Leveraging Lexical Cohesion and Disruption for Topic Segmentation
Anca-Roxana Şimon | Guillaume Gravier | Pascale Sébillot
Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing

pdf bib
A probabilistic segment model combining lexical cohesion and disruption for topic segmentation (Un modèle segmental probabiliste combinant cohésion lexicale et rupture lexicale pour la segmentation thématique) [in French]
Anca Simon | Guillaume Gravier | Pascale Sébillot
Proceedings of TALN 2013 (Volume 1: Long Papers)

2012

pdf bib
Automates lexico-phonétiques pour l’indexation et la recherche de segments de parole (Lexical-phonetic automata for spoken utterance indexing and retrieval) [in French]
Julien Fayolle | Fabienne Moreau | Christian Raymond | Guillaume Gravier
Proceedings of the Joint Conference JEP-TALN-RECITAL 2012, volume 1: JEP

pdf bib
The ETAPE corpus for the evaluation of speech-based TV content processing in the French language
Guillaume Gravier | Gilles Adda | Niklas Paulsson | Matthieu Carré | Aude Giraudel | Olivier Galibert
Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12)

The paper presents a comprehensive overview of existing data for the evaluation of spoken content processing in a multimedia framework for the French language. We focus on the ETAPE corpus which will be made publicly available by ELDA mid 2012, after completion of the evaluation campaign, and recall existing resources resulting from previous evaluation campaigns. The ETAPE corpus consists of 30 hours of TV and radio broadcasts, selected to cover a wide variety of topics and speaking styles, emphasizing spontaneous speech and multiple speaker areas.

2010

pdf bib
Utilisation de relations sémantiques pour améliorer la segmentation thématique de documents télévisuels
Camille Guinaudeau | Guillaume Gravier | Pascale Sébillot
Actes de la 17e conférence sur le Traitement Automatique des Langues Naturelles. Articles longs

Les méthodes de segmentation thématique exploitant une mesure de la cohésion lexicale peuvent être appliquées telles quelles à des transcriptions automatiques de programmes télévisuels. Cependant, elles sont moins efficaces dans ce contexte, ne prenant en compte ni les particularités des émissions TV, ni celles des transcriptions. Nous étudions ici l’apport de relations sémantiques pour rendre les techniques de segmentation thématique plus robustes. Nous proposons une méthode pour exploiter ces relations dans une mesure de la cohésion lexicale et montrons qu’elles permettent d’augmenter la F1-mesure de +1.97 et +11.83 sur deux corpus composés respectivement de 40h de journaux télévisés et de 40h d’émissions de reportage. Ces améliorations démontrent que les relations sémantiques peuvent rendre les méthodes de segmentation moins sensibles aux erreurs de transcription et au manque de répétitions constaté dans certaines émissions télévisées.

2008

pdf bib
On the Use of Web Resources and Natural Language Processing Techniques to Improve Automatic Speech Recognition Systems
Gwénolé Lecorvé | Guillaume Gravier | Pascale Sébillot
Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08)

Language models used in current automatic speech recognition systems are trained on general-purpose corpora and are therefore not relevant to transcribe spoken documents dealing with successive precise topics, such as long multimedia streams, frequently tacking reportages and debates. To overcome this problem, this paper shows that Web resources and natural language processing techniques can be effective to automatically adapt the baseline language model of an automatic speech recognition system to any encountered topic. More precisely, we detail how to characterize the topic of transcription segment and how to collect Web pages from which a topic-specific language model can be trained. Then, an adapted language model is obtained by combining the topic-specific language model with the general-purpose language model. Finally, new transcriptions are generated using the adapted language model and are compared with transcriptions previously obtained with the baseline language model. Experiments show that our topic adaptation technique leads to significant transcription quality gains.

pdf bib
Morphosyntactic Resources for Automatic Speech Recognition
Stéphane Huet | Guillaume Gravier | Pascale Sébillot
Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08)

Texts generated by automatic speech recognition (ASR) systems have some specificities, related to the idiosyncrasies of oral productions or the principles of ASR systems, that make them more difficult to exploit than more conventional natural language written texts. This paper aims at studying the interest of morphosyntactic information as a useful resource for ASR. We show the ability of automatic methods to tag outputs of ASR systems, by obtaining a tag accuracy similar for automatic transcriptions to the 95-98 % usually reported for written texts, such as newspapers. We also demonstrate experimentally that tagging is useful to improve the quality of transcriptions by using morphosyntactic information in a post-processing stage of speech decoding. Indeed, we obtain a significant decrease of the word error rate with experiments done on French broadcast news from the ESTER corpus; we also notice an improvement of the sentence error rate and observe that a significant number of agreement errors are corrected.

pdf bib
Un modèle multi-sources pour la segmentation en sujets de journaux radiophoniques
Stéphane Huet | Guillaume Gravier | Pascale Sébillot
Actes de la 15ème conférence sur le Traitement Automatique des Langues Naturelles. Articles longs

Nous présentons une méthode de segmentation de journaux radiophoniques en sujets, basée sur la prise en compte d’indices lexicaux, syntaxiques et acoustiques. Partant d’un modèle statistique existant de segmentation thématique, exploitant la notion de cohésion lexicale, nous étendons le formalisme pour y inclure des informations d’ordre syntaxique et acoustique. Les résultats expérimentaux montrent que le seul modèle de cohésion lexicale ne suffit pas pour le type de documents étudié en raison de la taille variable des segments et de l’absence d’un lien direct entre segment et thème. L’utilisation d’informations syntaxiques et acoustiques permet une amélioration substantielle de la segmentation obtenue.

2006

pdf bib
Corpus description of the ESTER Evaluation Campaign for the Rich Transcription of French Broadcast News
S. Galliano | E. Geoffrois | G. Gravier | J.-F. Bonastre | D. Mostefa | K. Choukri
Proceedings of the Fifth International Conference on Language Resources and Evaluation (LREC’06)

This paper presents the audio corpus developed in the framework of the ESTER evaluation campaign of French broadcast news transcription systems. This corpus includes 100 hours of manually annotated recordings and 1,677 hours of non transcribed data. The manual annotations include the detailed verbatim orthographic transcription, the speaker turns and identities, information about acoustic conditions, and name entities. Additional resources generated by automatic speech processing systems, such as phonetic alignments and word graphs, are also described.

2004

pdf bib
The ESTER Evaluation Campaign for the Rich Transcription of French Broadcast News
G. Gravier | J-F. Bonastre | E. Geoffrois | S. Galliano | K. McTait | K. Choukri
Proceedings of the Fourth International Conference on Language Resources and Evaluation (LREC’04)