Disparate Impact in Differential Privacy from Gradient Misalignment

2022/06/15 by Maria S. Esipova, Atiyeh Ashari Ghomi, Esipova, Maria S. +5 · 5 citations
Computer Science · Social Sciences · #Artificial Intelligence (cs.AI) #Cryptography and Security (cs.CR) #Ethics and Social Impacts of AI #FOS: Computer and information sciences #Machine Learning (cs.LG) #Privacy-Preserving Technologies in Data #Stochastic Gradient Optimization Techniques

paper · pdf · doi:10.48550/arxiv.2206.07737

openalex publication_date 2022/06/15 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

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.

Cited by

• FedPF: Accurate Target Privacy Preserving Federated Learning Balancing Fairness and Utility
• On Optimal Hyperparameters for Differentially Private Deep Transfer Learning
• SoftAdaClip: A Smooth Clipping Strategy for Fair and Private Model Training
• Private and Fair Machine Learning: Revisiting the Disparate Impact of Differentially Private SGD
• On the Fairness of Privacy Protection: Measuring and Mitigating the Disparity of Group Privacy Risks for Differentially Private Machine Learning

Related

• Removing Disparate Impact of Differentially Private Stochastic Gradient Descent on Model Accuracy
• Differential Privacy Has Disparate Impact on Model Accuracy
• Private and Fair Machine Learning: Revisiting the Disparate Impact of Differentially Private SGD
• Privacy at a Price: Exploring its Dual Impact on AI Fairness
• An Empirical Analysis of Fairness Notions under Differential Privacy