… we need fair learning. We show that the more biased the training data is, the higher the privacy
cost of achieving fairness for … empirical analysis for general machine learning algorithms. …

… differential privacy and fairness can be at odds with each other when we consider a learning
… In particular, we consider a simple binary classification setting where the learning algorithm …

… result highlighting the incompatibility of privacy and fairness. Section 4.3 talks about some
other directions we pursued along the trade-offs between privacy and fairness, and finally in …

… The focus on fairness in machine learning and its relationship to differential privacy was
explored in early work by the privacy community [7]. This work introduced the concept of treating …

… However, these mechanisms are inadequate for ensuring privacy and availability for fair …
fair data trading protocol in big data market to conduct privacy-preserving, availability and fair …

… Machine learning has changed our lives and will undoubtedly bring us more excitement.
However, its success is closely tied to the availability of large-scale training data, and as …

… ML from privacy, safety, and fairness perspectives, as illustrated in Figure 1:1 Privacy: The
speech-… user privacy, such as revealing sensitive information that people might want to keep …

… private deep learning. Specifically, we aim to study how different levels of imbalance in the
data affect the accuracy and the fairness of … We also study the fairness metrics demographic …

… requirements simultaneously in machine learning algorithms is under exploited. In this … one
classic machine learning model, logistic regression, and develop differentially private and fair …

… Among various privacy notions, differential privacy has become … private exponential
mechanism as a post-processing step to improve both fairness and privacy of supervised learning …