The 27th Workshop on the Semantics and Pragmatics of Dialogue


All MariLogue talks will be streamed online via MSTeams for conference participants only. Virtual presentations will be streamed live online.

Day 1 (16th August 2023) CET

09.30-10.00 CETRegistration
10.00-11.00 CETKeynote 1: Philippe Blache

Deep and compositional, or shallow and direct: two routes and one loop for a new approach to language understanding in conversation

Language understanding is a complex task, integrating different sources of information, from sounds and gestures to context. However, in spite of its complexity, this process is extremely fast and robust, performed in real time during conversations. Many studies have shown that this robustness and efficiency are made possible by different mechanisms: the ability to predict, the possibility of directly accessing entire pieces of meaning and the possibility to perform a “good-enough” processing, sufficient to access the meaning. These mechanisms, by substituting to the classical incremental and compositional architecture, facilitate the access to the meaning. However, existing models do not explain precisely when these facilitation mechanisms are triggered and whether they inhibit or on the contrary work in parallel with the standard ones.

I propose in this presentation a new model integrating both facilitation and standard mechanisms by revisiting the different stages of the processing: segmentation of the input, access to the corresponding meaning in long-term memory and integration to the interpretation under construction. This architecture is based on different features: unique representation of linguistic objects (independently from their granularities), control of the memory access (in particular thanks to search space reduction) and multiple-level prediction. This neuro-cognitive model provides a new framework explaining how deep and shallow mechanisms of language processing can cohabit. It is also a good candidate for explaining different effects of mismatch observed at the brain level.
11.00-11.30 CETBreak (coffee and pastry)
11.30-13.00 CETTalks Session 1

Joanna Kruyt, Katarína Polónyiová, Daniela Ostatníková and Stefan Benus:
Lexical entrainment on target words during task-oriented interaction in children with and without autism spectrum disorder

Carol Figueroa, Magalie Ochs and Gabriel Skantze:
Classification of Feedback Functions in Spoken Dialog Using Large Language Models and Prosodic Features

Ahmad Jabbar and Veda Kanamarlapudi:
Grounding with particles (VIRTUAL)
13.00-14.00 CETLunch
14.00-15.30 CETTalks Session 2

Josephine B. Fisher, Amelie S. Robrecht, Stefan Kopp and Katharina J. Rohlfing:
Exploring the Semantic Dialogue Patterns of Explanations — a Case Study of Game Explanations

Lina Mavrina and Stefan Kopp:
Predicting grounding state for adaptive explanation generation in analogical problem-solving

Jana Götze and David Schlangen:
Why Do You Say So? Dialogical Classification Explanations in the Wild and Elicited Through Classification Games
15.30-17.00 CETPoster session
Virtual posters will be recorded as 3-minute videos accompanied by brief explanations about the presented work. Participants can access these videos via the link provided in the schedule on the conference webpage. The authors of virtual posters are going to be available during the poster section via MSTeams.

Amandine Decker, Ellen Breitholtz, Christine Howes and Staffan Larsson:
Topic and genre in dialogue

Dan Green and Arash Eshghi:
One Signer at a Time? A Corpus Study of Turn-Taking Patterns in Signed Dialogue

Ryan Ka Yau Lai, Lu Liu, Haoran Yan and John W. Dubois
From position to function: Exploring word distributions within intonation units in American English conversation (VIRTUAL)

Claire To, Setareh Nasihati Gilani and David Traum:
Common Strategy Patterns of Persuasion in a Mission Critical and Time Sensitive Task (VIRTUAL)

Dominik Künkele and Simon Dobnik:
Referring as a collaborative process: learning to ground language through language games

Chiara Mazzocconi and Béatrice Priego-Valverde:
Humour in early interaction: what it can tell us about the child neuro-psychological, reasoning and linguistic development

Alex Lưu: Characterization of Discourse Salience in English Social Dialogs and its Application to Assessing Interactional Competence of Social Dialog Systems (VIRTUAL)

Ryan Ka Yau Lai:
rezonateR: An R package for analysing coherence in conversation (VIRTUAL)

Na Li and Robert Ross:
A Framework for Confusion Mitigation in Task-Oriented Interactions

Brielen Madureira and David Schlangen:
“Are you telling me to put glasses on the dog?” Content-Grounded Annotation of Instruction Clarification Requests in the CoDraw Dataset

Christine Howes, Vladislav Maraev and Ellen Breitholtz:
Menstruating vampires: What talk about taboos can tell us about dialogue

Darinka Verdonik, Simona Majhenič and Andreja Bizjak:
Are metadiscourse dialogue acts a category on their own?

Gayatri Purigilla and Radhika Mamidi:
Automatic classification of humor using Covid-19 related tweets (VIRTUAL).
17.00-17.30_CETSemDial Business Meeting (optional)
19.00-late CETDinner at the Novi svet pri Stolnici Restaurant

Day 2 (17th August 2023) CET

09.30-10.00 CETRegistration
10.00-11.00 CETKeynote 2: Liesbeth Degand

Discourse markers in interaction: distribution and functions

In this presentation I give an overview of the use of Discourse Markers (N= 1872) in a multi-genre corpus of spoken French. Contrasting their use in dialogic and monologic contexts, I will show how their distribution and function is impacted by their context of use. The focus will be on the relationship between frequency, polysemy and polyfunctionality of the markers, and how their (syntactic) form influences their functional distribution.
11.00-11.30 CETBreak (coffee and pastry)
11.30-13.00 CETTalks Session 3

Yikai Tseng, Takenobu Tokunaga and Hikaru Yokono:
Lexical Level Alignment in Dialogue

Arash Eshghi and Arash Ashrafzadeh:
Learning to generate and corr- uh I mean repair language in real-time (VIRTUAL)

Ryan Whetten, Enoch Levandovsky, Mir Tahsin Imtiaz and Casey Kennington:
Evaluating Automatic Speech Recognition and Natural Language Understanding in an Incremental Setting (VIRTUAL)
13.00-14.00 CETLunch
14.00-15.00 CETKeynote 3: Marko Robnik Šikonja

Challenges in explaining machine learning models for text

The area of Explainable Artificial Intelligence has developed many approaches for the explanation of machine learning models. The most successful methods are based on counterfactuals, prototypes, and perturbation of inputs. Unfortunately, none of these approaches work well to explain large language models, which currently dominate the natural language processing. The presentation will focus on challenges in using the most successful explanation methods for text classification, such as the interpretation of feature attributions, explanation of longer textual units, on-manifold vs. off-manifold explanations, unstable and uncertain explanations, and inadequate and unsystematic evaluation of explanation techniques. We will present possible solutions and outline a framework for more general explanation approaches.
15.00-15.30 CETBreak
15.30-17.00_CETTalks Session 4

Mathias Barthel, Henrike Helmer and Silke Reineke:
First users’ interactions with voice-controlled virtual assistants: A micro-longitudinal corpus study

Nicholas Walker, Stefan Ultes and Pierre Lison:
Retrieval-Augmented Neural Response Generation Using Logical Reasoning and Relevance Scoring

Mark Higger, Polina Rygina, Logan Daigler, Lara Bezarra, Zhao Han and Tom Williams:
Toward Open-World Human-Robot Interaction: What Types of Gestures Are Used in Task-Based Open-World Referential Communication?