On reconstructability of quadratic utility functions from the iterations in gradient methods

DOI10.1016/J.AUTOMATICA.2016.01.014MaRDI QIDQ254603zbMATH OpenOpenAlexFDO

Authors Farhad Farokhi, Iman Shames, Michael G. Rabbat, Mikael Johansson

Publication date 8 March 2016

Published in Automatica (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1509.05500

zbMATH Keywords

data confidentialitydata privacygradient methodsparameter identificationquadratic programmingstatistical inference

Mathematics Subject Classification ID

Quadratic programming (90C20) Filtering in stochastic control theory (93E11) Identification in stochastic control theory (93E12)

Abstract: In this paper, we consider a scenario where an eavesdropper can read the content of messages transmitted over a network. The nodes in the network are running a gradient algorithm to optimize a quadratic utility function where such a utility optimization is a part of a decision making process by an administrator. We are interested in understanding the conditions under which the eavesdropper can reconstruct the utility function or a scaled version of it and, as a result, gain insight into the decision-making process. We establish that if the parameter of the gradient algorithm, i.e.,~the step size, is chosen appropriately, the task of reconstruction becomes practically impossible for a class of Bayesian filters with uniform priors. We establish what step-size rules should be employed to ensure this.

Recommendations

• Encryption scheduling for remote state estimation under an operation constraint
• On the computational complexity of the secure state-reconstruction problem
• Distortion-based state security codes for distributed sensor networks
• Privacy preserving distributed optimization using homomorphic encryption
• Privacy-preserving dual stochastic push-sum algorithm for distributed constrained optimization

Cites work

• Convex programming in Hilbert space
• Differential Privacy: A Survey of Results
• Differentially private combinatorial optimization
• Differentially private empirical risk minimization
• scientific article; zbMATH DE number 951459 (Why is no real title available?)
• scientific article; zbMATH DE number 2107836 (Why is no real title available?)
• scientific article; zbMATH DE number 839070 (Why is no real title available?)
• scientific article; zbMATH DE number 926780 (Why is no real title available?)
• Numerical Optimization
• On reconstructability of quadratic utility functions from the iterations in gradient methods
• Persistence of excitation in extended least squares
• Privacy-preserving linear programming
• Set membership identification of nonlinear systems
• Set-membership identification of systems with parametric and nonparametric uncertainty

Cited in

(3)

• Modified Kalman filtering based multi-step-length gradient iterative algorithm for ARX models with random missing outputs
• On reconstructability of quadratic utility functions from the iterations in gradient methods
• Distributed safe resource allocation using barrier functions

This page was built for publication: On reconstructability of quadratic utility functions from the iterations in gradient methods

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q254603)