Coreference as an indicator of context scope in multimodal narrative

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Ilinykh, Nikolai, Lappin, Shalom, Sayeed, Asad, Loáiciga, Sharid
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866916801966768128
author Ilinykh, Nikolai
Lappin, Shalom
Sayeed, Asad
Loáiciga, Sharid
author_facet Ilinykh, Nikolai
Lappin, Shalom
Sayeed, Asad
Loáiciga, Sharid
contents We demonstrate that large multimodal language models differ substantially from humans in the distribution of coreferential expressions in a visual storytelling task. We introduce a number of metrics to quantify the characteristics of coreferential patterns in both human- and machine-written texts. Humans distribute coreferential expressions in a way that maintains consistency across texts and images, interleaving references to different entities in a highly varied way. Machines are less able to track mixed references, despite achieving perceived improvements in generation quality. Materials, metrics, and code for our study are available at https://github.com/GU-CLASP/coreference-context-scope.
format Preprint
id arxiv_https___arxiv_org_abs_2503_05298
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Coreference as an indicator of context scope in multimodal narrative
Ilinykh, Nikolai
Lappin, Shalom
Sayeed, Asad
Loáiciga, Sharid
Computation and Language
We demonstrate that large multimodal language models differ substantially from humans in the distribution of coreferential expressions in a visual storytelling task. We introduce a number of metrics to quantify the characteristics of coreferential patterns in both human- and machine-written texts. Humans distribute coreferential expressions in a way that maintains consistency across texts and images, interleaving references to different entities in a highly varied way. Machines are less able to track mixed references, despite achieving perceived improvements in generation quality. Materials, metrics, and code for our study are available at https://github.com/GU-CLASP/coreference-context-scope.
title Coreference as an indicator of context scope in multimodal narrative
topic Computation and Language
url https://arxiv.org/abs/2503.05298