Trade Privacy for Utility: A Learning-Based Privacy Pricing Game in Federated Learning
Authors:
- 28 May 2023
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 6307-6311
Abstract
To prevent implicit privacy disclosure in sharing gradients among data owners (DOs) under federated learning (FL), differential privacy (DP) and its variants have become a common practice to offer formal privacy guarantees with low overheads. However, individual DOs generally tend to inject larger DP noises for stronger privacy provisions (which entails severe degradation of model utility), while the curator (i.e., aggregation server) aims to minimize the overall effect of added random noises for satisfactory model performance. To address this conflicting goal, we propose a novel dynamic privacy pricing (DyPP) game which allows DOs to sell individual privacy (by lowering the scale of locally added DP noise) for differentiated economic compensations (offered by the curator), thereby enhancing FL model utility. Considering multi-dimensional information asymmetry among players (e.g., DO's data distribution and privacy preference, and curator's maximum affordable payment) as well as their varying private information in distinct FL tasks, it is hard to directly attain the Nash equilibrium of the mixed-strategy DyPP game. Alternatively, we devise a fast reinforcement learning algorithm with two layers to quickly learn the optimal mixed noise-saving strategy of DOs and the optimal mixed pricing strategy of the curator without prior knowledge of players' private information. Experiments on real datasets validate the feasibility and effectiveness of the proposed scheme in terms of faster convergence speed and enhanced FL model utility with lower payment costs.Keywords
All Related Versions
Funding Information
- NSFC (U22A2029,U20A20175)
- Fundamental Research Funds for the Central Universities
This publication has 17 references indexed in Scilit:
- Cost-Effective Federated Learning in Mobile Edge NetworksIEEE Journal on Selected Areas in Communications, 2021
- Pain-FL: Personalized Privacy-Preserving Incentive for Federated LearningIEEE Journal on Selected Areas in Communications, 2021
- Optimal Contract Design for Efficient Federated Learning With Multi-Dimensional Private InformationIEEE Journal on Selected Areas in Communications, 2020
- Hierarchical Incentive Mechanism Design for Federated Machine Learning in Mobile NetworksIEEE Internet of Things Journal, 2020
- Comprehensive Privacy Analysis of Deep Learning: Passive and Active White-box Inference Attacks against Centralized and Federated Learning2019
- Exploiting Unintended Feature Leakage in Collaborative Learning2019
- Deep Models Under the GAN2017
- Concentrated Differential Privacy: Simplifications, Extensions, and Lower Bounds2016
- Model Inversion Attacks that Exploit Confidence Information and Basic Countermeasures2015
- Privacy or Utility in Data Collection? A Contract Theoretic ApproachIEEE Journal of Selected Topics in Signal Processing, 2015