COPA
收藏资源简介:
displayName: COPA (Choice of Plausible Alternatives) labelTypes: - Multiple Choice license: - BSD 2-Clause mediaTypes: - Image - Text paperUrl: https://aclanthology.org/L18-1316.pdf publishDate: "2011-01-01" publishUrl: https://people.ict.usc.edu/~gordon/copa.html publisher: - University of Southern California tags: [] taskTypes: - Visual Question Answering --- ## 简介 似是而非的选择(COPA)评估为研究人员提供了一种评估开放领域常识因果推理进展的工具。 COPA 由 1000 个问题组成,平均分为开发和测试集,每组 500 个问题。每个问题都由一个前提和两个备选方案组成,其中的任务是选择更合理地与前提具有因果关系的备选方案。正确的选择是随机的,因此随机猜测的预期性能是 50%。 ## 引文 ``` @inproceedings{yeo2018visual, title={Visual choice of plausible alternatives: An evaluation of image-based commonsense causal reasoning}, author={Yeo, Jinyoung and Lee, Gyeongbok and Wang, Gengyu and Choi, Seungtaek and Cho, Hyunsouk and Amplayo, Reinald Kim and Hwang, Seung-won}, booktitle={Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018)}, year={2018} } ``` ## Download dataset :modelscope-code[]{type="git"}
显示名称:COPA(Choice of Plausible Alternatives,似是而非选择任务) labelTypes: - 多项选择(Multiple Choice) license: - BSD 2条款许可证(BSD 2-Clause) mediaTypes: - 图像(Image) - 文本(Text) paperUrl: https://aclanthology.org/L18-1316.pdf publishDate: "2011-01-01" publishUrl: https://people.ict.usc.edu/~gordon/copa.html publisher: - 南加州大学(University of Southern California) tags: [] taskTypes: - 视觉问答(Visual Question Answering) --- ## 简介 似是而非选择任务(Choice of Plausible Alternatives,简称COPA)评估集为研究者提供了一款用于评测开放域常识因果推理能力进展的标准化工具。该数据集共包含1000道问题,被平均划分为开发集与测试集,二者各含500道题目。每道题目均由一个前提陈述与两个备选选项构成,任务目标为选出与前提陈述具备更合理因果关联的备选选项。由于正确选项的位置是随机设置的,随机猜测的预期准确率为50%。 ## 引文 @inproceedings{yeo2018visual, title={Visual choice of plausible alternatives: An evaluation of image-based commonsense causal reasoning}, author={Yeo, Jinyoung and Lee, Gyeongbok and Wang, Gengyu and Choi, Seungtaek and Cho, Hyunsouk and Amplayo, Reinald Kim and Hwang, Seung-won}, booktitle={Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018)}, year={2018} } ## 下载数据集 :modelscope-code[]{type="git"}

- COPA数据集首次发表在《Proceedings of the 2011 Conference on Empirical Methods in Natural Language Processing》上,由Regina Barzilay和Lillian Lee提出,旨在解决因果关系理解问题。
- COPA数据集首次应用于自然语言处理领域的因果推理任务,成为评估模型因果关系理解能力的重要基准。
- COPA数据集在多个研究论文中被引用,进一步验证了其在因果关系理解任务中的有效性和挑战性。
- COPA数据集被纳入多个自然语言处理课程和教程,成为教学和研究的重要资源。
- COPA数据集在多个国际会议和研讨会上被广泛讨论,推动了因果关系理解研究的发展。
- 1Choice of Plausible Alternatives: An Evaluation of Commonsense Causal ReasoningAssociation for the Advancement of Artificial Intelligence · 2011年
- 2COPA-BERT: Enhancing Commonsense Causal Reasoning with BERTarXiv · 2020年
- 3Exploring the Limits of Transfer Learning with a Unified Text-to-Text TransformerGoogle Research · 2020年
- 4A Survey on Commonsense Reasoning: Advances and ChallengesarXiv · 2021年
- 5Improving Commonsense Reasoning with Multi-Task LearningarXiv · 2021年



