Local Differential Privacy in Decentralized Optimization
This work addresses privacy concerns for sensitive data in decentralized optimization, offering a method to enhance privacy with potential applications in distributed systems, though it appears incremental as it builds on existing optimization techniques.
The paper tackles the problem of ensuring local differential privacy in decentralized optimization by proposing a framework with random local aggregators that amplifies privacy by a constant, and demonstrates this with ADMM and gradient descent, showing experimental support for the theory.
Privacy concerns with sensitive data are receiving increasing attention. In this paper, we study local differential privacy (LDP) in interactive decentralized optimization. By constructing random local aggregators, we propose a framework to amplify LDP by a constant. We take Alternating Direction Method of Multipliers (ADMM), and decentralized gradient descent as two concrete examples, where experiments support our theory. In an asymptotic view, we address the following question: Under LDP, is it possible to design a distributed private minimizer for arbitrary closed convex constraints with utility loss not explicitly dependent on dimensionality? As an affiliated result, we also show that with merely linear secret sharing, information theoretic privacy is achievable for bounded colluding agents.