There has been an explosion of interest in fairness in machine learning. In large part, this has been motivated by societal issues highlighted in a string of well publicized cases such as gender biased job recommendation and racially biased criminal risk prediction algorithms. Both the recognition of the potential disparate impacts of machine learning due to historical bias in the data and the realization of how algorithmic decision making can exaggerate existing structural inequities has become increasingly well known.

中文翻译：

---

技术视角

人们对机器学习的公平性产生了浓厚的兴趣。在很大程度上，这是由一系列广为人知的案例中突出的社会问题所推动的，例如性别偏见的工作推荐和种族偏见的犯罪风险预测算法。由于数据中的历史偏差而对机器学习的潜在不同影响的认识以及算法决策如何夸大现有结构性不平等的认识都变得越来越广为人知。