The Price Of Privacy And The Limits Of Lp Decoding
STOC(2007)
摘要
This work is at the intersection of two lines of research. One line, initiated by Dinur and Nissim, investigates the price, in accuracy, of protecting privacy in a statistical database. The second, growing from an extensive literature on compressed sensing (see in particular the work of Donoho and collaborators [14, 7, 13, 11]) and explicitly connected to error-correcting codes by Candes and Tao, ([4]; see also [5, 3]), is in the use of linear programming for error correction.Our principal result is the discovery of a sharp threshold rho* approximate to 0.239, so that if rho < rho* and A is a random m x n encoding matrix of independently chosen standard Gaussians, where m = O(n), then with overwhelming probability over choice of A, for all x is an element of R-n, LP decoding corrects [rho m] arbitrary errors in the encoding Ax, while decoding can be made to fail if the error rate exceeds rho*. Our bound resolves an open question of Candes, Rudelson, Tao, and Vershyin [3] and (oddly, but explicably) refutes empirical conclusions of Donoho [11] and Candes et al [3]. By scaling and rounding we can easily transform these results to obtain polynomial-time decodable random linear codes with polynomial-sized alphabets tolerating any rho < rho* approximate to 0.239 fraction of arbitrary errors.In the context of privacy-preserving datamining our results say that any privacy mechanism, interactive or non-interactive, providing reasonably accurate answers to a 0.761 fraction of randomly generated weighted subset sum queries, and arbitrary answers on the remaining 0.239 fraction, is blatantly non-private.
更多查看译文
关键词
Privacy,LP Decoding,Compressed Sensing,Basis Pursuit
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络