Genomic privacy game
Script error: No such module "Draft topics".
Script error: No such module "AfC topic".
A genomic privacy game is a game theoretic model (see game theory) that is applied to solving a problem related to genomic privacy (or genetic privacy). Game theory, which tries to predict how the behavior of competitors influences the choices the other players make, can help data sharer find the best ways to share data while protecting the anonymity of the people contributing the data from hackers. In one study,[1] a two-player Stackelberg game model is applied to solving a membership inference attack targeting genomic data sharing. This game can be named a one-stage membership inference game (or OSMIG).[2] In another study,[3] a two-player Stackelberg game model is applied to solving a two-stage re-identification attack targeting genomic data sharing. This game can be named a multi-stage re-identification game (MSRIG).[4] In both of these two genomic privacy games, there are two players. Each player has his/her own strategy set and the payoff function can be calculated as a function of two players' strategies.
Game theoretic model[edit]
Players:
- Player 1: Defender (i.e., data sharer)
- Player 2: Attacker (i.e., data recipient)
Strategy sets:
- Player 1's strategy set: Player 1 needs to decide whether or not to mask (or redact) each attribute value for each record.
- Player 2's strategy set: Player 2 needs to decide whether or not to attack each record.
Payoff functions:
- Player 1's payoff is calculated as the value of the shared data minus the probability of success times the loss of Player 1 (i.e., the defender) for a successful attack, in which the value of shared data is a function of the defender's strategy, and the probability of success is a function of both players' strategies.
- Player 2's payoff is calculated as the probability of success times the gain of Player 2 (i.e., the attacker) for a successful attack minus the cost of the attack, in which the probability of success is a function of both players' strategies.
Note that, in a Stackelberg game, players moves sequentially. In this case, the leader (Player 1) moves first and the follower (Player 2) moves next.
In the OSMIG, the probability of success is calculated by simulating a one-stage membership inference attack.[5] In the MSRIG, the probability of success is calculated by simulating a two-stage re-identification attack.[6]
References[edit]
- ↑ Wan, Zhiyu; Vorobeychik, Yevgeniy; Xia, Weiyi; Clayton, Ellen Wright; Kantarcioglu, Murat; Malin, Bradley (2017-02-02). "Expanding Access to Large-Scale Genomic Data While Promoting Privacy: A Game Theoretic Approach". The American Journal of Human Genetics. 100 (2): 316–322. doi:10.1016/j.ajhg.2016.12.002. ISSN 0002-9297. PMC 5294764. PMID 28065469.
- ↑ Wan, Zhiyu (2017-12-30), GenoPriGas, retrieved 2021-12-18
- ↑ Wan, Zhiyu; Vorobeychik, Yevgeniy; Xia, Weiyi; Liu, Yongtai; Wooders, Myrna; Guo, Jia; Yin, Zhijun; Clayton, Ellen Wright; Kantarcioglu, Murat; Malin, Bradley A. (2021). "Using game theory to thwart multistage privacy intrusions when sharing data". Science Advances. 7 (50): eabe9986. Bibcode:2021SciA....7E9986W. doi:10.1126/sciadv.abe9986. PMC 8664254 Check
|pmc=value (help). PMID 34890225 Check|pmid=value (help). - ↑ Wan, Zhiyu (2021-10-01), zhywan/msrigs: Multi-Stage Re-Identification Game Solver, doi:10.5281/zenodo.5543369, retrieved 2021-12-18
- ↑ Sankararaman, Sriram; Obozinski, Guillaume; Jordan, Michael I.; Halperin, Eran (September 2009). "Genomic privacy and limits of individual detection in a pool". Nature Genetics. 41 (9): 965–967. doi:10.1038/ng.436. ISSN 1546-1718. PMID 19701190. Unknown parameter
|s2cid=ignored (help) - ↑ Gymrek, Melissa; McGuire, Amy L.; Golan, David; Halperin, Eran; Erlich, Yaniv (2013-01-18). "Identifying Personal Genomes by Surname Inference". Science. 339 (6117): 321–324. Bibcode:2013Sci...339..321G. doi:10.1126/science.1229566. PMID 23329047. Unknown parameter
|s2cid=ignored (help)
This article "Genomic privacy game" is from Wikipedia. The list of its authors can be seen in its historical and/or the page Edithistory:Genomic privacy game. Articles copied from Draft Namespace on Wikipedia could be seen on the Draft Namespace of Wikipedia and not main one.
