Diana Nicoleta Popa - From lexical towards contextualized meaning representation

14:00
Vendredi
27
Sep
2019
Organisé par : 
Diana-Nicoleta Popa
Intervenant : 
Diana-Nicoleta Popa
Équipes : 

Jury :

  • M. Eric GAUSSIER
    PREX, Laboratoire d'Informatique de Grenoble - Université Grenoble Alpes, directeur de thèse
  • M. Éric VILLEMONTE DE LA CLERGERIE
    Chargé de recherche, INRIA Paris, rapporteur
  • Mme Claire GARDENT
    Directrice de recherche, LORIA - Laboratoire Lorrain d'Informatique et ses Applications (CNRS), rapportrice
  • M. Laurent BESACIER
    Professeur, Laboratoire d'Informatique de Grenoble - Université Grenoble Alpes, examinateur
  • M. Alexandre ALLAUZEN
    Professeur, Université Paris-Sud, examinateur
  • M. James HENDERSON
    Chargé de Recherche, Idiap Research Institute et Université de Genève, examinateur
  • M. Julien PEREZ
    Senior Research Scientist, Naver Labs Europe, examinateur

 

Continuous word representations (word type embeddings) are at the basis of most modern natural language processing systems, providing competitive results particularly when input to deep learning models. However, important questions are raised concerning the challenges they face in dealing with complex natural language phenomena and regarding their ability to capture natural language variability. A first part of the thesis investigates a method for encoding complex phenomena such as entailment within a vector space by enforcing information inclusion. The second part of the thesis proposes a model for incorporating contextual knowledge into word representations by leveraging linguistic information.