distribution-over-quantities
收藏资源简介:
该数据集包含名词、动词和形容词的测量分布元组。它描述了如狮子体重、身高、人们通常吃早餐的时间以及音乐会时长等常识性问题的量化分布。数据集通过聚合测量和相应的提及值来表征大小、持续时间、重量等属性的量化分布。
This dataset comprises measurement distribution tuples of nouns, verbs, and adjectives. It delineates the quantitative distributions of common-sense queries such as the weight and height of lions, the usual times people have breakfast, and the duration of concerts. The dataset characterizes the quantitative distributions of attributes like size, duration, and weight by aggregating measurements and their corresponding mentioned values.
数据集概述
数据集名称
distribution-over-quantities
数据集内容
该数据集包含名词、动词和形容词的测量分布元组。数据集旨在描述如大小、持续时间、重量等属性的量化分布,通过聚合测量和相应的提及值来实现。
数据格式
每个数据条目包含以下字段:
obj: 对象head: 头部dim: 维度mean: 均值perc_5: 5th 百分位测量值perc_25: 25th 百分位测量值median: 中位数perc_75: 75th 百分位测量值perc_95: 95th 百分位测量值std: 标准差
引用信息
若在您的研究中使用或讨论此数据集,请引用以下论文:
@InProceedings{DOQ, title = {How Large Are Lions? Inducing Distributions over Quantitative Attributes}, author = "Yanai Elazar, Abhijit Mahabal, Deepak Ramachandran, Tania Bedrax-Weiss, Dan Roth}, booktitle = {Proceedings of ACL 2019}, year = {2019}




