arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2009.06190cs.LGstat.ML
收藏

半监督学习中的公平性约束

Fairness Constraints in Semi-supervised Learning

University of Technology Sydney(悉尼科技大学);University of Illinois at Chicago(伊利诺伊大学芝加哥分校)

Tao Zhang, Tianqing Zhu, Mengde Han, Jing Li, Wanlei Zhou, Philip S. Yu

摘要:

机器学习中的公平性受到了相当大的关注。然而,大多数关于公平学习的研究集中在监督学习或无监督学习。很少有人考虑半监督设置。然而,在现实中,大多数机器学习任务依赖于包含标记和未标记数据的大型数据集。公平学习的关键问题之一是公平性和准确性之间的平衡。以前的研究认为,增加训练集的大小可以有更好的权衡。我们相信,用未标记的数据增加训练集可以达到类似的结果。因此,我们开发了一个公平的半监督学习,这是制定为一个优化问题的框架。这包括用于优化准确性的分类器损失、用于优化未标记数据预测的标记传播损失以及用于优化公平性水平的对标记和未标记数据的公平性约束。该框架进行逻辑回归和支持向量机下的公平性度量的不同的影响和不同的虐待。从理论上分析了半监督学习中的偏差、方差和噪声分解等鉴别力的来源。大量实验表明,该方法能够实现公平的半监督学习,并且比公平的监督学习在准确性和公平性之间取得了更好的折衷。

英文摘要:

Fairness in machine learning has received considerable attention. However, most studies on fair learning focus on either supervised learning or unsupervised learning. Very few consider semi-supervised settings. Yet, in reality, most machine learning tasks rely on large datasets that contain both labeled and unlabeled data. One of key issues with fair learning is the balance between fairness and accuracy. Previous studies arguing that increasing the size of the training set can have a better trade-off. We believe that increasing the training set with unlabeled data may achieve the similar result. Hence, we develop a framework for fair semi-supervised learning, which is formulated as an optimization problem. This includes classifier loss to optimize accuracy, label propagation loss to optimize unlabled data prediction, and fairness constraints over labeled and unlabeled data to optimize the fairness level. The framework is conducted in logistic regression and support vector machines under the fairness metrics of disparate impact and disparate mistreatment. We theoretically analyze the source of discrimination in semi-supervised learning via bias, variance and noise decomposition. Extensive experiments show that our method is able to achieve fair semi-supervised learning, and reach a better trade-off between accuracy and fairness than fair supervised learning.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑