Towards Multimodal Metaphor Understanding: A Chinese Dataset and Model for Metaphor Mapping Identification

Dongyu Zhang, Shengcheng Yin, Jingwei Yu, Zhiyao Wu, Zhen Li, Chengpei Xu, Xiaoxia Wang, Feng Xia · 2025

Read Towards Multimodal Metaphor Understanding: A Chinese Dataset and Model for Metaphor Mapping Identification by Dongyu Zhang, Shengcheng Yin, Jingwei Yu, Zhiyao Wu, Zhen Li, Chengpei Xu, Xiaoxia Wang, Feng Xia (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-02434.txt.

Description

Dongyu Zhang and co-authors built this 2025 resource for teaching machines to read metaphors in Chinese ads. A metaphor maps one domain onto another, like calling time a thief. Computers can often spot that a metaphor is present, yet naming its two domains is harder. Non-English, image-based material for this task was almost missing.

Their data set, CM3D, collects Chinese ads that mix words and pictures, with the source and target domains labeled. Alongside it, they propose a model that mimics human step-by-step reasoning to find the mapping. A two-level training scheme steadies the search. The work opens metaphor research beyond English text and toward real ads, where meaning rides on image and slogan together.

Catalog record

CREATORDongyu Zhang, Shengcheng Yin, Jingwei Yu, Zhiyao Wu, Zhen Li, Chengpei Xu, Xiaoxia Wang, Feng Xia
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE63 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-02434

Browse the full library · About the Tower of Babel