Fairness by Learning Orthogonal Disentangled Representations
Fairness by Learning Orthogonal Disentangled Representations
复制标题
DOI：
10.1007/978-3-030-58526-6_44
复制
发表时间：
2020-03
期刊：
影响因子：
--
通讯作者：
Mhd Hasan Sarhan;N. Navab;Abouzar Eslami;Shadi Albarqouni
中科院分区：
文献类型：
--
作者：
Mhd Hasan Sarhan;N. Navab;Abouzar Eslami;Shadi Albarqouni
学习有区别的强有力表示是机器学习系统的关键一步。在特定任务上表现良好的同时引入对任意讨厌或敏感属性的不变性是表示学习中的一个重要问题。这主要是通过从学习的表示中清除敏感信息来实现的。在本文中，我们提出了一种新的去纠缠不变表示问题的方法。我们解开有意义的和敏感的表示，强制执行正交约束作为独立性的代理。我们明确地强制有意义的表示是不可知的敏感信息的熵最大化。所提出的方法进行了评估五个公开可用的数据集和比较的最先进的方法学习的公平性和不变性实现的最先进的性能在三个数据集和可比的性能在其余的。此外，我们进行了烧蚀研究，以评估每个组件的效果。
Learning discriminative powerful representations is a crucial step for machine learning systems. Introducing invariance against arbitrary nuisance or sensitive attributes while performing well on specific tasks is an important problem in representation learning. This is mostly approached by purging the sensitive information from learned representations. In this paper, we propose a novel disentanglement approach to invariant representation problem. We disentangle the meaningful and sensitive representations by enforcing orthogonality constraints as a proxy for independence. We explicitly enforce the meaningful representation to be agnostic to sensitive information by entropy maximization. The proposed approach is evaluated on five publicly available datasets and compared with state of the art methods for learning fairness and invariance achieving the state of the art performance on three datasets and comparable performance on the rest. Further, we perform an ablative study to evaluate the effect of each component.