TY - JOUR
T1 - The value of privacy
T2 - Strategic data subjects, incentive mechanisms, and fundamental limits
AU - Wang, Weina
AU - Ying, Lei
AU - Zhang, Junshan
N1 - Funding Information:
This work was supported in part by the National Science Foundation under grants ECCS-1255425 and SaTC-1618768. Authors’ addresses: W. Wang, L. Ying, and J. Zhang, Goldwater Center, School of ECEE, Arizona State University, Tempe, AZ, 85281; emails: {weina.wang, lei.ying.2, junshan.zhang}@asu.edu. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from permissions@acm.org. © 2018 ACM 2167-8375/2018/08-ART8 $15.00 https://doi.org/10.1145/3232863
Funding Information:
This work was supported in part by the National Science Foundation under grants ECCS-1255425 and SaTC-1618768.
Publisher Copyright:
© 2018 ACM
PY - 2018/8
Y1 - 2018/8
N2 - We study the value of data privacy in a game-theoretic model of trading private data, where a data collector purchases private data from strategic data subjects (individuals) through an incentive mechanism. One primary goal of the data collector is to learn some desired information from the elicited data. Specifically, this information is modeled by an underlying state, and the private data of each individual represents his of her knowledge about the state. Departing from most of the existing work on privacy-aware surveys, our model does not assume the data collector to be trustworthy. Further, an individual takes full control of his or her own data privacy and reports only a privacy-preserving version of his or her data. In this article, the value of ϵ units of privacy is measured by the minimum payment among all nonnegative payment mechanisms, under which an individual's best response at a Nash equilibrium is to report his or her data in an ϵ-locally differentially private manner. The higher ϵ is, the less private the reported data is. We derive lower and upper bounds on the value of privacy that are asymptotically tight as the number of data subjects becomes large. Specifically, the lower bound assures that it is impossible to use a lower payment to buy ϵ units of privacy, and the upper bound is given by an achievable payment mechanism that we design. Based on these fundamental limits, we further derive lower and upper bounds on the minimum total payment for the data collector to achieve a given accuracy target for learning the underlying state and show that the total payment of the designed mechanism is at most one individual's payment away from the minimum.
AB - We study the value of data privacy in a game-theoretic model of trading private data, where a data collector purchases private data from strategic data subjects (individuals) through an incentive mechanism. One primary goal of the data collector is to learn some desired information from the elicited data. Specifically, this information is modeled by an underlying state, and the private data of each individual represents his of her knowledge about the state. Departing from most of the existing work on privacy-aware surveys, our model does not assume the data collector to be trustworthy. Further, an individual takes full control of his or her own data privacy and reports only a privacy-preserving version of his or her data. In this article, the value of ϵ units of privacy is measured by the minimum payment among all nonnegative payment mechanisms, under which an individual's best response at a Nash equilibrium is to report his or her data in an ϵ-locally differentially private manner. The higher ϵ is, the less private the reported data is. We derive lower and upper bounds on the value of privacy that are asymptotically tight as the number of data subjects becomes large. Specifically, the lower bound assures that it is impossible to use a lower payment to buy ϵ units of privacy, and the upper bound is given by an achievable payment mechanism that we design. Based on these fundamental limits, we further derive lower and upper bounds on the minimum total payment for the data collector to achieve a given accuracy target for learning the underlying state and show that the total payment of the designed mechanism is at most one individual's payment away from the minimum.
KW - Data collection
KW - Differential privacy
KW - Randomized response
UR - http://www.scopus.com/inward/record.url?scp=85053392911&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85053392911&partnerID=8YFLogxK
U2 - 10.1145/3232863
DO - 10.1145/3232863
M3 - Article
AN - SCOPUS:85053392911
SN - 2167-8375
VL - 6
JO - ACM Transactions on Economics and Computation
JF - ACM Transactions on Economics and Computation
IS - 2
M1 - 8
ER -