Casual Conversations v2
收藏arXiv2022-11-11 更新2024-07-31 收录
下载链接：
https://ai.facebook.com/blog/shedding-light-on-fairness-in-ai-with-a-new-data-set
下载链接
链接失效反馈官方服务：
资源简介：
Casual Conversations v2是由Meta AI与香港科技大学合作设计的大型同意驱动数据集，旨在测量AI算法的偏见和鲁棒性。该数据集包含全面的标签类别，如年龄、性别、语言/方言、地理位置、残疾、身体装饰和属性等，以确保模型评估的全面性和公平性。数据集的创建过程考虑了全球多样性和地区差异，旨在解决AI模型在不同人群中的性能不均等问题，推动AI技术的公平和包容性发展。
Casual Conversations v2 is a large consent-driven dataset co-designed by Meta AI and The Hong Kong University of Science and Technology, which aims to measure the bias and robustness of AI algorithms. The dataset includes comprehensive annotation categories such as age, gender, language/dialect, geographic location, disability, body adornments and attributes, to ensure comprehensive and fair evaluation of AI models. The development of the dataset takes into account global diversity and regional differences, targeting to address the problem of uneven performance of AI models across different demographic groups, and promote the equitable and inclusive development of AI technology.
提供机构：
Meta AI
创建时间：
2022-11-11
