ConMeC
收藏资源简介:
ConMeC(普通名词隐喻语料库)是由辛辛那提大学创建的一个新的隐喻分辨率数据集,包含6000个从维基百科中提取的句子。每个句子都配有一个目标普通名词,并由人类标注者标注该普通名词在上下文中是否被隐喻使用。该数据集是迄今为止最大的专注于普通名词隐喻分辨率的数据集,旨在为隐喻分辨率任务提供基准,并评估NLP管道在理解隐含语言方面的能力。
ConMeC (Corpus of Common Noun Metaphors) is a novel metaphor resolution dataset developed by the University of Cincinnati. It contains 6,000 sentences extracted from Wikipedia, with each sentence paired with a target common noun. Human annotators have labeled whether each target common noun is used metaphorically in its contextual setting. This is the largest dataset dedicated to common noun metaphor resolution to date, and its core objectives are to provide a benchmark for the metaphor resolution task and evaluate the ability of NLP pipelines to understand implicit language.

- 1ConMeC: A Dataset for Metonymy Resolution with Common Nouns辛辛那提大学 · 2025年



