{"pk":49856,"title":"Locating strongly informative utterances in conversation using multimodal cues","subtitle":null,"abstract":"Interaction theories argue that mutual understanding between\nspeakers in natural conversations arises from building shared\nknowledge (common ground), but no model specifies what\ninformation is retained or under what conditions. Previous\nstudies have used Information Theory metrics to quantify the\ndynamics of information exchanged between participants but\nlack an efficient way to identify which information becomes\ncommon ground. These attempts furthermore limited them-\nselves to the study of conversation transcripts, overlooking\nnonverbal cues like visuals and intonation. To address this,\nwe propose a method for annotating new corpora using models\ntrained on a subset of annotated utterances. Results show a fair\napplicability (κ ≃0.3) across corpora, though this is strongly\nmodulated by the conversational task being investigated.","language":"eng","license":{"name":"","short_name":"","text":null,"url":""},"keywords":[{"word":"Discourse; Natural Language Processing; Computer-based experiment; Neural Networks"}],"section":"Papers with Poster Presentation","is_remote":true,"remote_url":"https://escholarship.org/uc/item/2k16f7qf","frozenauthors":[{"first_name":"Eliot","middle_name":"","last_name":"Ma‘s","name_suffix":"","institution":"Aix Marseille University","department":""},{"first_name":"Philippe","middle_name":"","last_name":"Blache","name_suffix":"","institution":"ILCB","department":""},{"first_name":"Leonor","middle_name":"","last_name":"Becerra-Bonache","name_suffix":"","institution":"Aix-Marseille University","department":""}],"date_submitted":null,"date_accepted":null,"date_published":"2025-01-01T10:00:00-08:00","render_galley":null,"galleys":[{"label":"PDF","type":"pdf","path":"https://journalpub.escholarship.org/cognitivesciencesociety/article/49856/galley/37818/download/"}]}