Domain Adaptation meets Individual Fairness. And they get along
Domain Adaptation meets Individual Fairness. And they get along
复制标题
DOI：
10.48550/arxiv.2205.00504
复制
发表时间：
2022-05
期刊：
影响因子：
--
通讯作者：
Debarghya Mukherjee;Felix Petersen;M. Yurochkin;Yuekai Sun
中科院分区：
文献类型：
--
作者：
Debarghya Mukherjee;Felix Petersen;M. Yurochkin;Yuekai Sun
许多算法偏差的例子是由分布偏移引起的。例如，机器学习（ML）模型在训练数据中代表性不足的人口统计群体上的表现往往更差。在本文中，我们利用算法公平性和分布偏移之间的这种联系来表明算法公平性干预可以帮助ML模型克服分布偏移，并且域自适应方法（用于克服分布偏移）可以减轻算法偏差。特别是，我们证明了（i）强制执行适当的个体公平性（IF）概念可以在协变量移位假设下提高ML模型的分布外准确性，以及（ii）有可能调整表示对齐方法进行域适应以强制个体公平性。前者是意料之外的，因为在制定综合框架干预措施时没有考虑到分配的变化。后者也是出乎意料的，因为代表对齐不是一个共同的方法，在个人公平文献。
Many instances of algorithmic bias are caused by distributional shifts. For example, machine learning (ML) models often perform worse on demographic groups that are underrepresented in the training data. In this paper, we leverage this connection between algorithmic fairness and distribution shifts to show that algorithmic fairness interventions can help ML models overcome distribution shifts, and that domain adaptation methods (for overcoming distribution shifts) can mitigate algorithmic biases. In particular, we show that (i) enforcing suitable notions of individual fairness (IF) can improve the out-of-distribution accuracy of ML models under the covariate shift assumption and that (ii) it is possible to adapt representation alignment methods for domain adaptation to enforce individual fairness. The former is unexpected because IF interventions were not developed with distribution shifts in mind. The latter is also unexpected because representation alignment is not a common approach in the individual fairness literature.