Modelling features relevant for discourse deixis resolution in dialogue

Home  ›  Thesis Topics  ›  Ongoing Theses/Internships  ›  Modelling features relevant for discourse deixis resolution in dialogue
Modelling features relevant for discourse deixis resolution in dialogue

Research questions: 

  • Discourse deixis mention identification: What distinguishes discourse deixis and basic anaphora anaphors? What role does the semantics of the predicate play? Obviously, the identification of the markable as a discourse deixis one also depends on the previous context (antecedent) – can we model the context as a feature? And how to attend to the most relevant parts of it? 
  • Discourse deixis mention resolution: How do humans find antecedents for discourse deixis anaphors? If the antecedent consists of several sentences/utterances, what exactly links them together? Why does this piece of discourse stand out? How important are discourse relations?  How can features relevant for discourse deixis resolution be modelled? Some statistics of potential findings would also be beneficial. 
  • Are there any crucial differences between discourse deixis resolution in text and dialogue?  
  • Can we improve/extend existing model(s) for discourse deixis resolution by adding the features discussed above? 

Literature suggestions: 

  1. Discourse Deixis and Discourse Processing, B. Webber, 1988 
  1. Structure and Ostension in the Interpretation of Discourse Deixis, B. Webber, 1991 
  1. Fully Automatic Resolution of ‘it’, ‘this’, and ‘that’ in Unrestricted Multi-Party Dialog, M. Müller, 2008 
  1. Resolving pronominal reference to abstract entities, D. Byron, 2002 
  1. Abstract anaphora resolution in Danish, C. Navarretta, 2000 
  1. Dialogue acts, synchronizing units, and anaphora resolution, M. Eckert and M. Strube, 2000 
  1. A machine learning approach to pronoun resolution in spoken dialogue, M. Strube and M. Müller, 2003 
  1. A Mention-Ranking Model for Abstract Anaphora Resolution, A. Marasovic et al., 2017  
  1. Neural Anaphora Resolution in Dialogue, H. Kobayashi et al., 2021 
  1. Anaphora Resolution in Dialogue: Description of the DFKI-TalkingRobots System for the CODI-CRAC 2021 Shared-Task, T. Anikina et al., 2021 
  1. Higher-order coreference resolution with coarse-to-fine inference, K. Lee et al., 2018 
  1. The CODI-CRAC 2021 shared task on anaphora, bridging, and discourse deixis in dialogue. S. Khosla et al., 2021