As machine learning becomes more widespread throughout society, aspects including data privacy and fairness must be carefully considered, and are crucial for deployment in highly regulated industries. Unfortunately, the application of privacy enhancing technologies can worsen unfair tendencies in models. In particular, one of the most widely used techniques for private model training, differentially private stochastic gradient descent (DPSGD), frequently intensifies disparate impact on groups within data. In this work we study the fine-grained causes of unfairness in DPSGD and identify gradient misalignment due to inequitable gradient clipping as the most significant source. This observation leads us to a new method for reducing unfairness by preventing gradient misalignment in DPSGD.

翻译：随着机器学习在社会各领域的普及，数据隐私与公平性等议题必须审慎考量，这对高度监管行业的部署至关重要。不幸的是，隐私增强技术的应用可能加剧模型中的不公平倾向。具体而言，最常用的私有模型训练技术之一——差分隐私随机梯度下降（DPSGD）——常会加剧对数据中群体的差异性影响。本研究深入剖析DPSGD不公平性的细粒度成因，发现由非公平梯度裁剪导致的梯度对齐偏差是最主要的根源。这一发现促使我们提出一种新方法，通过防止DPSGD中的梯度对齐偏差来减少不公平性。