Invited Talk: Gabriella Lapesa on 'NLP-supported (e-)deliberation: interdisciplinary challenges and real-world applications'

We like to invite you to our upcoming invited talk by Dr. Gabriella Lapesa from the University of Stuttgart. The talk will take place on Thursday, June 29, 2023 at 16:00, as part of our graduate seminar in Room 3408-1630 at Appelstraße 9A:

Speaker

Dr. Gabriella Lapesa

University of Stuttgart

https://www.ims.uni-stuttgart.de/institut/team/Lapesa/

Time and location

Thursday, June 29, 16:00, 3408-1630 (Appelstr. 9A)

Title

NLP-supported (e-)deliberation: interdisciplinary challenges and real-world applications

Abstract

Digital innovations have reshaped citizen participation both in terms of the amount of people that can be involved and in terms of the forms of their interaction. How about the quality of the discussions, which are not just a preparatory step, but at the core any democratic decision making process? Discussion quality is not guaranteed to scale up to large user crowds — quite the opposite. In this talk, I will discuss the interdisciplinary challenges related to the use of NLP to support e-deliberation (i.e., group decision making in online forums), with a focus on the crucial role played by human moderators. The question of “what is a good contribution to a discussion” is crucial in both NLP/Argument Mining, which has produced theoretical definitions for argument quality (and large-scale annotated resources which represent the empirical counterpart of the theories), and in the Social Sciences/Deliberative Theory, with theories of deliberative quality for certain aspects complementary to the Argument Mining ones (and annotated resources that are typically of high quality but small size). Can we bridge these theories and the annotations they produced, to support human moderators? Beyond that, what are the challenges for the implementation of NLP-supported moderation in the “real world”?

External participants (i.e. neither students nor employees of LUH) are kindly requested to send a short, informal email to office@ai.uni-hannover.de to register