arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1906.02589cs.LGcs.AIstat.ML
收藏

通过解纠缠实现灵活公平的表示学习

Flexibly Fair Representation Learning by Disentanglement

University of Toronto(多伦多大学);Vector Institute(向量研究所);University of Tübingen(蒂宾根大学);Google Research(谷歌研究院)

Elliot Creager, David Madras, Jörn-Henrik Jacobsen, Marissa A. Weis, Kevin Swersky, Toniann Pitassi, Richard Zemel

摘要:

我们考虑的问题,学习表示,实现组和子组的公平性方面的多个敏感属性。从解开表示学习文献的灵感,我们提出了一种算法,用于学习数据集的紧凑表示,这是有用的重建和预测,但也是\emph{flexibly fair},这意味着他们可以很容易地在测试时修改,以实现子组人口统计奇偶性方面的多个敏感属性及其连接。我们的经验表明,由此产生的编码器-这不需要敏感的属性进行推理-使适应一个单一的表示,以各种公平的分类任务与新的目标标签和子群定义。

英文摘要:

We consider the problem of learning representations that achieve group and subgroup fairness with respect to multiple sensitive attributes. Taking inspiration from the disentangled representation learning literature, we propose an algorithm for learning compact representations of datasets that are useful for reconstruction and prediction, but are also \emph{flexibly fair}, meaning they can be easily modified at test time to achieve subgroup demographic parity with respect to multiple sensitive attributes and their conjunctions. We show empirically that the resulting encoder---which does not require the sensitive attributes for inference---enables the adaptation of a single representation to a variety of fair classification tasks with new target labels and subgroup definitions.

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。

↑