Proxy Non-Discrimination in Data-Driven Systems
Proxy Non-Discrimination in Data-Driven Systems
复制标题
数据驱动系统中的代理非歧视
DOI：
--
复制
发表时间：
2017
期刊：
影响因子：
--
通讯作者：
S. Sen
中科院分区：
文献类型：
--
作者：
Anupam Datta;Matt Fredrikson;Gihyuk Ko;Piotr (Peter) Mardziel;S. Sen
机器学习系统从训练数据中继承了对受保护类别（即历史上受歧视的群体）的偏见。通常，这些偏见并不明显，它们依赖于训练算法发现的微妙相关性，因此难以察觉。我们将数据驱动系统中的代理歧视（一类表明偏见的属性）形式化为存在对系统输出有因果影响的受保护类别相关性。我们在一组社会数据集上评估了一种实现方式，展示了如何根据这些属性验证系统，并在出现违规情况时进行修复。
Machine learnt systems inherit biases against protected classes, historically disparaged groups, from training data. Usually, these biases are not explicit, they rely on subtle correlations discovered by training algorithms, and are therefore difficult to detect. We formalize proxy discrimination in data-driven systems, a class of properties indicative of bias, as the presence of protected class correlates that have causal influence on the system's output. We evaluate an implementation on a corpus of social datasets, demonstrating how to validate systems against these properties and to repair violations where they occur.