Chinese Multimodal Metaphor Mapping Dataset (CM3D)
收藏资源简介:
CM3D数据集是由大连理工大学等机构开发的中文多模态隐喻映射数据集,包含6108个文本-图像对,主要来源于中文广告。该数据集通过标注目标域和源域的隐喻表达,旨在促进非英语语言中的隐喻理解研究。数据集的内容涵盖了广告中的隐喻映射,提供了详细的注释,帮助研究人员深入理解隐喻的认知机制。数据集的创建过程包括从广告中提取隐喻表达,并进行人工标注。该数据集的应用领域主要集中在自然语言处理中的隐喻理解,特别是多模态隐喻的识别与解释,旨在解决隐喻映射识别中的复杂问题。
The CM3D dataset is a Chinese multimodal metaphor mapping dataset developed by institutions such as Dalian University of Technology. It comprises 6108 text-image pairs primarily sourced from Chinese advertisements. By annotating metaphorical expressions in both the target domain and source domain, this dataset is designed to facilitate research on metaphor comprehension in non-English languages. The dataset covers metaphorical mappings in advertisements, and provides detailed annotations to help researchers gain in-depth insights into the cognitive mechanisms of metaphor. The creation process of the dataset involves extracting metaphorical expressions from advertisements and conducting manual annotation. Its main application fields focus on metaphor comprehension in natural language processing, especially the recognition and interpretation of multimodal metaphors, aiming to address complex issues in metaphor mapping recognition.

- 1Towards Multimodal Metaphor Understanding: A Chinese Dataset and Model for Metaphor Mapping Identification大连理工大学, 澳门大学, 关西大学, 新南威尔士大学, 皇家墨尔本理工大学 · 2025年



