Lien pour suivre les présentations orales à distance:
https://univ-grenoble-alpes-fr.zoom.us/j/94499179700?pwd=WEtHK0llaWZmTWtmTnhPUlVqRmlGUT09
6 décembre :
9h00-10h00 - Accueil+café
10h00-10h30 - Introduction
- Présentation de la conférence (Benjamin Lecouteux) [pdf]
- Présentation du GdR LIFT (Claire Gardent)
10h30-11h45 - Invité : Laurent Besacier
Titre : Self-Supervised Learning for Low Resource Speech Tasks
Résumé : Self-supervised learning using huge unlabeled data has been successfully explored for image processing and natural language processing. Since 2019, recent works also investigated self-supervised representation learning from speech. They were notably successful to improve performance on downstream tasks such as speech recognition. These recent works suggest that it is possible to reduce dependence on labeled data for building speech systems through acoustic representation learning. In this talk I will present an overview of these recent approaches to self-supervised learning from speech and show my own investigations to use them in spoken language processing tasks for which size of training data is limited.
12h00-13h30 - Pause déjeuner+café
13h30-14h30 - Invitée : Alda Mari
Titre : Epistemic future in questions, MICA and evidence quality
Résumé : In our talk we address the question of the interpretation of Italian MICA in questions with future in Italian. Contrary to recent views according to which MICA is a common ground management operator, we propose that MICA is a metaevaluator that ranks possibilities as low in a normality scale and ultimately gives rise to an inference of surprise in questions with future.
This analysis is cast in a framework where future is analyzed as an epistemic modal (Giannakidou and Mari 2018; in opposition to views according to which future is an evidential (Mari 2010, Eckardt and Beltrama 2019, Frana and Menendez-Benito 2019) and in which the modal skeleton contains a metaevaluator. We propose a distinction between evidence *source* and evidence *quality* that allows us to ground the epistemic use of future in degraded evidence leaving room for the uncertainty presupposition over which MICA operates.
14h30-15h30 - Contributions ; présentations orales
- Le projet ANR Autogramm et l'extraction automatique de grammaire - Illustration par la négation. Sylvain Kahane, Bruno Guillaume, Kim Gerdes, Bernard Caron, Sylvain Loiseau. [pdf]
- Analyse orientée corpus d'universaux de Greenberg sur Universal Dependencies. Hee-Soo Choi, Bruno Guillaume, Karën Fort. [pdf]
15h30-16h00 - Pause café
16h00-17h30 - Contributions ; posters
- Convertir le Trésor de la Langue Française en Ontolex-Lemon: un zeste de données liées
Sina Ahmadi, Mathieu Constant, Karën Fort, Bruno Guillaume, John P. McCrae. [pdf]
- Deux corpus audio transcrits de langues rares (japhug et na) normalisés en vue d'expériences en traitement du signal
Benjamin Galliot, Guillaume Wisniewski, Séverine Guillaume, Laurent Besacier, Guillaume Jacques, Alexis Michaud, Solange Rossato, Minh-Châu Nguyên, Maxime Fily. [pdf]
- Exploration de systèmes end-to-end pour la reconnaissance automatique de la parole spontanée
Solène Evain. [pdf]
- Reading interlinearized glossed texts: inference of linguistic features from free translations
Sylvain Loiseau. [pdf]
- Segmentation en mots faiblement supervisée pour la documentation automatique des langues
Shu Okabe, François Yvon, Laurent Besacier. [pdf]
- Simplification syntaxique de textes à base de représentations sémantiques en DMRS
Hijazi Rita. [pdf]
7 décembre :
8h30-9h00 - Accueil+café
9h00-10h00 - Invité : Daan van Esch
Titre : Building Language Technology for Everyone
Résumé : [TBA]
10h00-10h30 - Contribution ; présentation orale
- Le(s)? chinois du Shun-pao 申報. Pierre Magistry. [pdf]
10h30-11h00 - Pause café
11h00-12h00 - Contributions ; présentations orales
- Spécialisation de modèles neuronaux pour la transcription phonémique : premiers pas vers la reconnaissance de mots pour les langues rares. Cécile Macaire, Guillaume Wisniewski, Séverine Guillaume, Benjamin Galliot, Guillaume Jacques, Alexis Michaud, Solange Rossato, Minh-Châu Nguyên, Maxime Fily. [pdf]
- DinG -- a corpus of transcriptions of real-life, oral, spontaneous multi-party dialogues between French-speaking players of Catan. Maria Boritchev, Maxime Amblard. [pdf]
12h00-13h30 - Pause déjeuner+café
13h30-14h30 - Invité : Matti Miestamo
Titre : Negation - Language typology, documentation and description
Résumé : Cross-linguistic typological work on negation has paid most attention to standard negation, i.e. the negation of declarative verbal main clauses (Dahl 1979; Payne 1985; Dryer 2013a,b,c; Miestamo 2005, 2013). Other aspects of negation that have received at least some attention in large-scale typological studies include the negation of imperatives (van der Auwera & Lejeune 2013), the negation of stative (nonverbal, existential, etc.) predications (Croft 1991; Eriksen 2011; Veselinova 2013), the negation of indefinite pronouns (Haspelmath 1997, 2013; Van Alsenoy 2014), abessives (Stolz et al. 2007), the effects of negation on the marking of NPs (Miestamo 2014), and negative replies to questions (Holmberg 2015) – for an overview of typological work on negation, see Miestamo 2017. Currently, typological work is underway on various aspects of the typology of negation: e.g., Veselinova’s work on negative lexicalizations and the relationship between negation and TAM, Miestamo & Koptjevskaja Tamm’s work on antonyms, Van Olmen’s work on negative imperatives, and Mauri & Sansò's work on anticircumstantial clauses as well as Miestamo, Shagal & Silvennoinen’s work on negation in dependent clauses. In this talk, I will give an overview of current typological knowledge of negation, focusing especially on interesting issues arising from my own work on the domain. I will also show how typological knowledge on negation can be made use of in language documentation and description, presenting a typologically and functionally oriented questionnaire for describing the domain of negation (Miestamo 2016) and giving an update on how the questionnaire is being used in an ongoing project describing negation in a number of languages from around the world.
14h30-15h15 - Contribution ; présentation orale
- On avertives as complex negative event descriptions. Patrick Caudal. [pdf]
- Quelques exemples de négation dans les langues créoles. Emmanuel Schang. [pdf]
15h15-16h00 - Conclusion / café
