Identification des reprises et interprétation automatique des expressions pronominales dans des textes en français
Summary
This doctoral thesis (Université Blaise-Pascal, GRIL, with Xerox Research Centre Europe, defended December 2001) examines how French texts link a later expression back to an earlier one ("reprises") and how pronouns can be interpreted automatically. Part I sets out a typology of such links: shared denotation, shared description, "membre-de" (member-of), "distingué-de" (distinguished-from), and reference to a discourse. It sets associative anaphora, deixis and date references apart from these. It then defines evaluation criteria that align key and response reference chains and score them by recall, precision, substitution, over-generation and under-generation. These criteria are compared with the link-based scheme of Vilain et al., Popescu-Belis's exclusive core classes and the B-3 algorithm. An inter-annotator experiment tests which notions annotators apply consistently. Part II builds a pronoun resolution system on morphological analysis and the XIP parser. Its methodology covers machine implementation, a corpus study and statistical checking of its hypotheses. The work matters because it supplies a corpus-grounded annotation and evaluation framework for French anaphora. The provided excerpt does not include the system's final scores.
Key points
- The typology is compared with the MATE coreference annotation scheme, Halliday and Hasan's cohesion model and Salmon-Alt's referential annotation.
- The operationality test used three texts, one expert and five annotators, with mean, variance, majority opinion and an operationality threshold as measures.
- The thesis reviews prior approaches: binding theory, selection restrictions, centering and veins theory, plus the systems of Hobbs, Lappin & Leass, Kennedy & Boguraev, CogNIAC and Popescu-Belis.
- The resolver relies on a frozen, non-final version of XIP, whose main weakness is prepositional-phrase attachment; an appendix shows these errors on a Paribas news text.
- A separate program converts the annotated files to XML. The thesis is distributed under CC BY-NC-SA 4.0 (HAL tel-01152394).
Cite
@inproceedings{trouilleux2001identification,
title = {{Identification des reprises et interprétation automatique des expressions pronominales dans des textes en français}},
author = {Trouilleux, François},
booktitle = {Thèse, Université Blaise-Pascal, Clermont-Ferrand},
year = {2001},
url = {https://hal.science/tel-01152394}
}References
- [1℄ Salah Aït-Mokhtar et Jean-Pierre Chanod . In remental nite-state Pro eedings of the 5th Conferen e on Applied Natural Language Pro essing (ANLP-97) , Washington, DC, USA, 1997. parsing. In [2℄ Salah Aït-Mokhtar et Jean-Pierre Chanod . Subje t and obje t de-ACL Workshop on Automati Information E… find
- ACL'93 , p. 156163, Ohio State University, États-Unis, 1993. find
- [5℄ Amit Bagga et Bre k Baldwin . Pro eedings of the LREC'98 Workshop on Linguisti Coreferen e , hains. In Grenade, Espagne, 1998. find
- CogNIAC : A Dis ourse Pro essing Engine . Thèse de [6℄ Bre k Baldwin . do torat, Fa ulty of Engineering and Applied S ien e, University of Pennsylvania, 1995. find
- [7℄ Bre k Baldwin . CogNIAC : High pre ision Pro eedings of the ACL'97/EACL'97 Workshop on Operational fa tors in pra ti al, robust anaphora resolution , p. 38 ledge and linguisti resour es. In 45, Madrid, Espagne, 1997. find
- [8℄ C t lina Barbu et Ruslan Mitkov . Evaluation tool for rule-based anaphora resolution methods. In Pro eedings of the 39th Annual Meeting of the Asso iation for Computational Linguisti s (ACL'2001) , p. 3441, Toulouse, Fran e, 2001. find our card
- Pro-eedings of the Se ond Colloquium on Dis ourse Anaphora and Anaphor Resolution (DAARC2) , p. 3442, Lan aster University, Royaume-Uni, 1998. find
- [9℄ Mi hael Barlow . Feature mismat hes and anaphora resolution. In Pro eedings of In Algorithms for s oring oreferen e oreferen e with limited know-Re her hes sur le français parlé , p. 237358. Publi ations de l'Université de Proven e, 1999. find
- [10℄ Gabriel G. Bès . La phrase verbale noyau en français. In [11℄ Gabriel G. Bès . Empiri ité en linguistique et grammaire de Montague : la sémantique en 5P et la ompositionalité. GRIL, Université Blaise-Pas al, avril 2001. find
- [12℄ Simon Botley et Anthony M. M Enery , dire teurs de publi ation. Corpus-based and Computational Approa hes to Dis ourse Anaphora , pitre Dis ourse Anaphora : The need for synthesis. John Benjamins, Amsterdam, 2000. find
- [13℄ Susan E. Brennan , Marilyn W. Friedman et Carl J. Pollard . A Pro eedings of the 25th Annual Meeting of the Asso iation for Computational Linguisti s (ACL'87) , p. 155162, entering approa h to pronouns. find
- Stanford, Californie, États-Unis, 1987. find
- [14℄ Donna K. Byron . The Un ommon Denominator. guisti s - Spe ial Issue on Anaphora Resolution , Á paraître. Disponible à l'adresse : http://www.s.ro hester.edu/u/dbyron/papers.html . [15℄ Donna K. Byron et Joel R. Tetreault . A exible ar hite ture for refe-Pro eedings of the 9th Conferen e of th… ↗
- [16℄ Jaime G. Carbonnel et Ralf G. Brown . Anaphora resolution : A multi-Pro eedings of the 12th International Conferen e on Computational Linguisti s (COLING'88) , p. 96101, Budapest, 1998. find
- [17℄ David M. Carter . Interpreting anaphors in natural language texts . Ellis strategy approa h. In Horwood, Royaume-Uni, 1987. find
- Langue naturelle et logique : La sémantique intensionnelle de Ri hard Montague . Peter Lang, [18℄ Mi hel Chambreuil et Jean-Claude Pariente . find
- [19℄ Mi hel Charolles et Catherine S hnedeker . Coréféren e et identité. Le problème des référents évolutifs. [20℄ Nan y A. Chin hor et Beth Sundheim . Message Understanding Confe-Pro eedings of the AAAI Spring Symposium on Empiri al Methods in Dis ourse Interpretation and Generation , Stanford, 199… find
- [21℄ Noam Chomsky . Le tures on Government and Binding . Foris Publi ations, ren e (MUC) Tests of Dis ourse Pro essing. Dordre ht, 1981. find
- [22℄ Dennis Connolly , John D. Burger et David S. Day . A ma hine learning approa h to anaphori Somers , dire teurs de publi ation, p. 133144. UCL Press, Londres, 1997. ha-In Computational Lin-Langages , 112, p. 106126, 1993. find
- In referen e. In Daniel B. Jones et Harold L. New Methods in Language Pro essing , Les formes de reprise dans le dis ours. Anaphore et haînes de référen e . Presses Universitaires de Rennes, 1995. find
- [24℄ Fran is Corblin . Celui- i anaphorique : un mentionnel. Langue française , [23℄ Fran is Corblin . 120, p. 3343, 1998. find
- [25℄ Dan Cristea , Nan y Ide et Laurent Romary . Veins theory : A model of global dis ourse ohesion and ACL'98 , p. 281285, Montréal, Canada, 1998. find
- [26℄ Ido Dagan et Alon Itai . Automati Pro eedings of the 13th International Conferen e on Computational Linguisti s (COLING'90) , p. 330332, Hel-resolution of anaphora referen es. In sinki, 1990. find
- [27℄ Lauren e Danlos . Event oreferen e in Bouillon et Federi a Busa , dire teurs de publi ation, Word Meaning , p. 216241. Cambridge University Press, 2001. find
- [28℄ Sarah Davies et Massimo Poesio . http://www.ogs i.ed.a .uk/~poesio/MATE/ oreferen e.html , 1998. ↗
- [29℄ Barbara Di Eugenio . On the usage of kappa to evaluate agreement on oding tasks. In Pro eedings of the Se ond International Conferen e on Language Resour es and Evaluation , Athènes, Grè e, 2000. find
- [30℄ David R. Dowty , Robert E. Wall et Stanley Peters . Introdu tion to Montague Semanti s . D. Reidel Publishing Company, Dordre ht, Holland, 1981. find
- Di tionnaire en y lopédique des s ien es du langage . Éditions du Seuil, Paris, 1972. find
- [31℄ Oswald Du rot et Tzvetan Todorov . [32℄ Antonio Ferrández , Manuel Palomar et Lidia Moreno . Anaphora resolution in unrestri ted texts with partial parsing. COLING-ACL'98 , p. 385391, Montréal, Canada, 1998. find
- [33℄ Steve Fligelstone . Developing a s heme for annotating text to show anaphori relations. In G. Leitner , dire teur de publi ation, in Corpus Linguisti s , p. 153170. Mouton de Gruyter, Berlin, 1992. find
- [34℄ Gottlob Frege . É rits logiques et philosophiques . Éditions du Seuil, Paris, 1971. find
- [35℄ Niyu Ge , John Hale et Eugene Charniak . Pro eedings of the 6th Workshop on Very Large anaphora resolution. In Corpora , p. 161170, Montréal, Canada, 1998. find
- [36℄ Gilles-Gaston Granger . La véri ation . Éditions Odile Ja ob, Paris, 1992. find
- Le bon usage. Grammaire française. Refondue par [37℄ Mauri e Grevisse . André Goosse. Du ulot, Paris - Louvain-la-Neuve, treizième édition, 1993. find
- Pro eedings of COLING-oheren e. In pro essing of large orpora for the ausal dis ourses. In Pierrette The Language of Coding S hemes for Co-referen e. Pro eedings of In New Dire tions A statisti al approa h to [38℄ Ralph Grishman et Beth Sundheim , dire teurs de publi ation. dings of the Sixth Messag… find
- [39℄ Barbara J. Grosz , Aravind K. Joshi et S ott Weinstein . Centering, a framework for modeling the lo al Linguisti s , 21(2), p. 175204, 1995. find
- [40℄ Barbara J. Grosz et Canda e L. Sidner . Lost intuitions and forgotten intentions. In Marilyn A. Walker , Aravind K. Joshi et Ellen F. Prin e , dire teurs de publi ation, 51. Oxford University Press, New York, 1998. find
- [41℄ Gsi-Erli . Le di tionnaire AlethDi . Version 1.5.4, 15 dé embre 1994. find
- [42℄ Franz Günthner et Hubert Lehmann . Rules for pronominalisation. In Pro eedings of the First Conferen e of the European Chapter of the Asso iation for Computational Linguisti s , p. 144151, Pise, Italie, 1983. find
- [43℄ M. A. K. Halliday et Ruqaiya Hasan . Cohesion in English . Longman, 1976. find
- [44℄ Lynette Hirshman et Nan y Chin hor . MUC-7 Coreferen e Task De-nition. Version 3.0. In Pro eedings of the Seventh Message Understanding Conferen e (MUC-7) , http://www.mu.sai . om , 1998. S ien e Appli ations International Corporation. ↗
- [45℄ Jerry Hobbs . Resolving pronoun referen es. [46℄ Nan y Ide et Dan Cristea . A hierar hi al a Pro eedings of ACL-2000 , p. 416424, Hong Kong, 2000. find
- [47℄ Kerstin Jonasson . Le nom propre. Constru tions et interprétations. Du-bility. In ulot, Louvain-la-Neuve, 1994. find
- [48℄ Megumi Kameyama . Intrasential Walker , Aravind K. Joshi et Ellen F. Prin e , dire teurs de publi a-Centering Theory in Dis ourse , p. 89112. Oxford University Press, tion, New York, 1998. find
- [49℄ Hans Kamp , Di k Crou h et Josef van Genabith . for Computational Semanti s (FRACAS). Deliverable D2. Spe i ation of Coverage. http://www. Linguisti [50℄ Hans Kamp et Uwe Reyle . find
- [51℄ Lauri Karttunen . Dis ourse referents. Syntax and Semanti s. Notes from the Linguisti re teur de publi ation, Underground , p. 363385. A ademi [52℄ Andrew Kehler , John Bear et Douglas Appelt . The Need for A - urate Alignment in Natural Language System Evaluation. Linguisti s , 27(2), 2001. find
- Pro ee-Morgan Computational oheren e of dis ourse. Centering Theory in Dis ourse , p. 39 Lingua , 44, p. 311338, 1976. ount of referential a essi-entering : A ase study. In Marilyn A. find
- A Framework ogs i.ed.a .uk/~fra as/ , 1994. find
- From Dis ourse to Logi . Kluwer A ademi In James D. M Cawley , di-Press, New York, 1976. find
- Computational [53℄ Christopher Kennedy et Branimir Boguraev . In Pro eedings of the 16th International Conferen e on Computational Linguisti s (CO-LING'96) , Copenhague, Danemark, 1996. ryone : Pronominal anaphora resolution without a parser. find
- [54℄ Georges Kleiber , Catherine S hnedeker et Lauren e Ujma . L'ana-on eption l'autre. In L'anaphore asso iative. Re-her hes linguistiques XIX , p. 564. Université de Metz, 1994. phore asso iative, d'une [55℄ Shalom Lappin et Herbert J. Leass . An algorithm for pronominal ana-Computational Linguis… find
- [56℄ William C. Mann et Sandra A. Thompson . theory : Toward a fun tional theory of text organization. 281, 1988. find
- [57℄ Daniel Mar u , Estibaliz Amorrortu et Magdalena Romera . Ex-onstru ting a orpus of dis ourse trees. In Pro eedings of the ACL'99 Workshop on Standards and Tools for Dis ourse Tagging , p. 4857, periments in Maryland, États-Unis, 1999. find
- [58℄ Patri io Martínez-Bar o , Rafael Muñoz , Saliha Azzam , Manuel Palomar et Antonio Ferrández . Evaluation of pronoun resolution algorithm for spanish dialogues. In Università Ca' Fos ari, Venise, Italie, 1999. find
- [59℄ MATE Multilevel Annotation, Tools Engineering. Telemati s Proje t LE4-8370. http://mate.nis.sdu.dk , 2000. ↗
- [60℄ Ruslan Mitkov . Robust pronoun resolution with limited knowledge. In Pro eedings of COLING-ACL'98 , p. 869875, Montréal, Canada, 1998. find our card
- [61℄ Ruslan Mitkov . Anaphora resolution : The state of the art. paper (Based on the COLING/ACL'98 tutorial on anaphora resolution). http://www.wlv.a.uk/sles/ ompling/papers/mitkov-99a.pdf , University of Wolverhampton, Royaume-Uni, 1999. ↗
- [62℄ Ruslan Mitkov . Towards more solution. In Pro eedings of the Se ond International Conferen e on Language Resour es and Evaluation , p. 13091314, Athènes, Grè e, 2000. find
- [63℄ Ruslan Mitkov , Lamia Belguith et Malgorzata Stys . Multilingual Pro eedings of the Third International Conferen e on Empiri al Methods in Natural Language Pro essing (EMNLP-3) , robust anaphora resolution. In p. 716, Grenade, Espagne, 1998. find
- Pronominalisation et ea ement du sujet en génération automatique de textes en langues romanes . Thèse de do torat, Université [64℄ Fiammetta Namer . find
- Anaphora for eve-Rhethori al stru ture Text , 8(3), p. 243 Pro eedings of VEXTAL'99 , p. 325332, Working omprehensive evaluation in anaphora re-[65℄ Tetsuya Nasukama . Robust method of pronoun resolution using full-Pro eedings of the 15th Conferen e on Computational Linguisti s (COLING'94) , p. 11… find
- [68℄ Jesús Peral , Manuel Palomar et Antonio Ferrández . Coreferen e-oriented interlingual slot stru ture and ma hine translation. In Pro eedings of the ACL Workshop on Coreferen e and Its Appli ations , p. 6976, University of Maryland, États-Unis, 1999. find
- [69℄ Massimo Poesio . MATE Annotation Guidelines Coreferen e. 2000. http://www.ims.uni-stuttgart.de/projekte/mate/mdag/ r/ r_1.html . ↗
- [70℄ Massimo Poesio et Renata Vieira . A Computational Linguisti s , 24(2), p. 183216, 1998. find
- nite des ritpion use. [71℄ Andrei Popes u-Belis . Évaluation numérique de la résolution de la référen e : ritiques et propositions. 40(2), p. 117142, 1999. find
- Modélisation multi-agent des é hanges langa-giers : appli ation au problème de la référen e et à son évaluation . Thèse de [72℄ Andrei Popes u-Belis . do torat, Université Paris XI, Orsay, 1999. find
- [73℄ Andrei Popes u-Belis et Isabelle Robba . Cooperation between pro-Pro eedings of the ACL'97/EACL'97 Workshop on Operational fa tors in pra ti al, robust anaphora resolution , p. 9499, Madrid, Espagne, 1997. noun and referen e resolution for unrestri ted texts. In [74℄ Andrei Popes u-Belis et Is… find
- Pro eedings of the LREC'98 Workshop on Linguisti Coreferen e , Granada, Spain, 1998. find
- [75℄ Aarne Ranta . Type-Theoreti al Grammar . Oxford University Press, Ox-evaluating referen e resolution. In ford, 1994. find
- [76℄ Anne Reboul , Balkanski , Cé ile Gaiffe , Andrei Popes u-Belis , Isabelle Robba , Laurent Romary et Gérard Sabah . Le projet CERVICAL : Représentations mentales, référen e aux objets et aux événements. Rapport te hnique, CRIN et LIMSI, 1997. find
- Anaphora and Semanti Interpretation . Croom Helm, [77℄ Tanya Reinhart . find
- Londres, 1983. o-referen e anno-orpus-based investigation of de-Traitement Automatique des Langues , Three new methods for Briffault , Xavier Bertrand [78℄ Tanya Reinhart . Coreferen e and bound anaphora : A restatement of the Linguisti s and Philosophy , 6(1), p. 4788, 1983. anaphora questions. find
- [79℄ Monique Rolbert . Résolution de formes pronominales dans l'interfa e d'interrogation d'une base de données. s ien es de Luminy., 1989. find
- [80℄ Maximiliano Saiz-Noeda et Manuel Palomar . Semanti In Natural Language Pro essing NLP 2000, Se ond International Conferen e , p. 204 driven method to solve pronominal anaphora in spanish texts. 221, Patras, Grè e, 2000. Springer. find
- [81℄ Susanne Salmon-Alt . Du Traitement Automatique des Langues , 42(2), 2001. find
- [82℄ Susanne Salmon-Alt . Référen e et dialogue nalisé : de la linguistique à un modèle opérationnel . Thèse de do torat, Université Henri-Poin aré, tielle. Nan y 1, 2001. find
- [83℄ Canda e L. Sidner . Fo using for interpretation of pronouns. tional Linguisti s , 7(4), p. 217231, 1981. find
- [84℄ Mi hael Strube et Udo Hahn . Fun tional rential oheren e in information stru ture. p. 309344, 1999. find
- [85℄ Roland Stu kardt . Robust anaphor resolution : Design and evaluation of the ROSANA system. In Pro eedings of ROMAND 2000 : Workshop on Robust Methods in Analysis of Natural Language Data , p. 4357, Lausanne, 2000. find
- Sele ted papers of Ri hard Montague . Yale University Press, New Haven, 1974. find
- [86℄ Ri hmond H. Thomason , dire teur de publi ation. [87℄ François Trouilleux . Identi ation et propres dans des textes en français. Blaise-Pas al, Clermont-Ferrand, 1997. find
- [88℄ François Trouilleux , Gabriel G. Bès et Éri luation of inter-annotator agreement in the observation of anaphori referential relations. In Pro eedings of the Third International Conferen e on Dis ourse Anaphora and Anaphor Resolution (DAARC2000) , Lan aster University, Royaume-Uni, 2000. find
- [89℄ François Trouilleux , Éri nen . Coreferen e resolution evaluation based on des riptive spe i ity. In Pro eedings of the Se ond International Conferen e on Language Resour es and Evaluation , Athènes, Grè e, 2000. find
- [90℄ Agnès Tutin , Trouilleux , François Gaussier , Annie Zaenen , Stéphanie Rayot et Georges Antoniadis . Thèse de do torat, Fa ulté des knowledge-orpus à la théorie : l'annotation ( o-)référen-Computa-entering Grounding refe-Computational Linguisti s , 25(3), lassement automatique des noms Mémoi… find
- Annotating a large [91℄ Mar Vilain , John Burger , John Aberdeen , Dennis Connolly et Lynette Hirshman . A model-theoreti Pro eedings of the Sixth Message Understanding Conferen e (MUC-6) , p. 45 52, San Fran is o, 1995. Morgan Kaufmann. find
- [92℄ Marilyn A. Walker . Evaluating dis ourse pro essing algorithms. In eedings of ACL'89 , p. 251261, University of British Columbia, Canada, 1989. find
- [93℄ Marilyn A. Walker , Aravind K. Joshi et Ellen F. Prin e , dire teurs Centering Theory in Dis ourse . Oxford University Press, de publi ation. find
- Preferen e Semanti s . Stanford AI Laboratory memo AIM-[94℄ Yori k Wilks . 206, Stanford University, 1973. find
- Understanding Natural Language . [95℄ Terry Winograd . find
- L'anaphore et les pronoms. Une introdu tion à la syntaxe générative. Presses Universitaires du Septentrion, Villeneuve-d'As q, [96℄ Anne Zribi-Hertz . 1996. oreferen e s oring s heme. find
This page in Markdown → https://md.research.elamur.ai/free/en/papers/trouilleux-2001-identification-reprises-interpretation-automatique/