Saved in:
Bibliographic Details
Main Authors: Chang, Kent K., Cramer, Mackenzie Hanh, Ho, Anna, Nguyen, Ti Ti, Yuan, Yilin, Bamman, David
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2505.17536
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910003466600448
author Chang, Kent K.
Cramer, Mackenzie Hanh
Ho, Anna
Nguyen, Ti Ti
Yuan, Yilin
Bamman, David
author_facet Chang, Kent K.
Cramer, Mackenzie Hanh
Ho, Anna
Nguyen, Ti Ti
Yuan, Yilin
Bamman, David
contents While multimodal large language models (LLMs) excel at dialogue, whether they can adequately parse the structure of conversation -- conversational roles and threading -- remains underexplored. In this work, we introduce a suite of tasks and release TV-MMPC, a new annotated dataset, for multimodal conversation structure understanding. Our evaluation reveals that while all multimodal LLMs outperform our heuristic baseline, even the best-performing model we consider experiences a substantial drop in performance when character identities of the conversation are anonymized. Beyond evaluation, we carry out a sociolinguistic analysis of 350,842 utterances in TVQA. We find that while female characters initiate conversations at rates in proportion to their speaking time, they are 1.2 times more likely than men to be cast as an addressee or side-participant, and the presence of side-participants shifts the conversational register from personal to social.
format Preprint
id arxiv_https___arxiv_org_abs_2505_17536
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Multimodal Conversation Structure Understanding
Chang, Kent K.
Cramer, Mackenzie Hanh
Ho, Anna
Nguyen, Ti Ti
Yuan, Yilin
Bamman, David
Computation and Language
While multimodal large language models (LLMs) excel at dialogue, whether they can adequately parse the structure of conversation -- conversational roles and threading -- remains underexplored. In this work, we introduce a suite of tasks and release TV-MMPC, a new annotated dataset, for multimodal conversation structure understanding. Our evaluation reveals that while all multimodal LLMs outperform our heuristic baseline, even the best-performing model we consider experiences a substantial drop in performance when character identities of the conversation are anonymized. Beyond evaluation, we carry out a sociolinguistic analysis of 350,842 utterances in TVQA. We find that while female characters initiate conversations at rates in proportion to their speaking time, they are 1.2 times more likely than men to be cast as an addressee or side-participant, and the presence of side-participants shifts the conversational register from personal to social.
title Multimodal Conversation Structure Understanding
topic Computation and Language
url https://arxiv.org/abs/2505.17536