Project Title: Conflicting Interpretations in Dialogue
TL;DR: Investigating misunderstandings in collaborative dialogues: identification, recovery, and model performance
Description
Misunderstandings in communication, especially within collaborative dialogue settings, present significant challenges in Natural Language Processing (NLP). This project explores the phenomenon of conflicting interpretations among interlocutors, specifically focusing on how misunderstandings arise and how they can be computationally modeled within collaborative communication scenarios. Current NLP methodologies often assume a single, definitive interpretation for reference expressions; however, this research investigates situations where speakers and listeners assign different referents to the same expressions—a common yet underexplored phenomenon in both human-human and human-AI interactions.
Our empirical analysis currently centers on dialogues from the MapTask dataset (Anderson et al., 1991), in which participants collaborate using asymmetric map information, creating significant potential for misunderstandings. Reference expressions in these dialogues are being systematically annotated to capture various types of interpretations, facilitating a deeper understanding and enabling refinement of computational models for reference resolution and grounding processes.
To further advance our research, we also plan to design experiments explicitly aimed at producing collaborative dialogues prone to misunderstandings. The resulting annotated datasets will support the development of computational methods capable of identifying misunderstandings, modeling their resolution processes, and facilitating recovery strategies. By testing Large Language Models (LLMs) within these collaborative dialogue contexts, we aim to enhance their ability to recognize and address misunderstandings. This project will thus contribute annotated datasets and actionable computational models, advancing both practical applications of LLMs in collaborative communication and our theoretical understanding of incremental mutual understanding.
The project is being carried out by Nan Li. He is a Ph.D. candidate in the NLP group at the Department of Information and Computing Sciences at Utrecht University, supervised by Prof. Massimo Poesio and Prof. Albert Gatt. For more information, please contact Nan at n.li[AT]uu.nl.
References
Anderson, A. H., Bader, M., Bard, E. G., Boyle, E., Doherty, G., Garrod, S., Isard, S., Kowtko, J., McAllister, J., Miller, J., Sotillo, C., Thompson, H. S., & Weinert, R. (1991). The Hcrc Map Task Corpus. Language and Speech, 34(4), 351–366.
Ginzburg, J. (2012). The Interactive Stance. Oxford University Press.