Skip to content

Privacy-preserving Truth Discovery (PPTD)

什么是 PPTD ？

Goal: find truthful information among conflicting data based on each user's weight.

  • If a data item provided by a user is closer to the aggregated result, this user will be assigned a higher weight
  • If a user holds a higher weight, the data of this user will be counted more in the execution

相关工作

  • LLG145，truth discovery 算法
  • MJS151，第一个提出解决 data privacy
  • XLL192，在个别用户 offline 的情况下保护隐私
  • TWY183，使用 GC，假设两个服务器不共谋
  • LMS184，使用差分隐私

[XLX20] - V-PATD

Catch you if you deceive me: Verifiable and privacy-aware truth discovery in crowdsensing systems

(ASIA CCS'20, October 5-9, 2020, Taipei, Taiwan)

"The first verifiable and privacy-aware truth discovery protocol in crowdsensing systems."

扩充：增量真值发现

与以往的真值发现方法不同，增量真值发现方法不涉及迭代阶段，并且是从真值更新开始的，然后在观察到新对象时更新每个用户的权重值。

用户集合表示为 U=\{u_1,u_2,\ldots, u_k\}，对象集合表示为 O=\{o_1,o_2,\ldots,o_M\}，w_m^k 表示为系统内第 k 个用户对于对象 o_m 的权重。

增量真值发现的目标就是计算当前时间戳对象 o_m 的基础真值 x_m。在增量真值发现的过程中，一个时间段内每个用户只感知一个对象，并在下一个时间内重新感知到新对象。

假设当前时间戳用户观测到的对象为 o_m，也就是说之前感知到的对象为 o_1,o_2,\ldots, o_{m-1}，那么增量真值发现的过程包含下面两个过程：

  • 真值估计：服务器根据上一个时间戳中计算出的权重 \{w_{m-1}^k\}_{k=1}^K 来估计对象的真值：
x_m = \frac{\sum_{k=1}^K w_{m-1}^K x_m^k}{\sum_{k=1}^K w_{m-1}^k}
  • 权重估计：服务器根据用户 u_k 的感知数据和真值的差来计算该用户在此刻对于对象 o_m 的权重。

思考：会不会有滞后性？

  1. C. Miao, W. Jiang, L. Su, Y. Li, S. Guo, Z. Qin, H. Xiao, J. Gao, and K. Ren. Cloud-Enabled Privacy-Preserving Truth Discovery in Crowd Sensing Systems. In Proceedings of the ACM SenSys. ACM, 183–196. 2015. ↩

  2. G. Xu, H. Li, S. Liu, M. Wen, and R. Lu. Efficient and Privacy-Preserving Truth Discovery in Mobile Crowd Sensing Systems. IEEE Transactions on Vehicular Technology 68, 4 (2019), 3854–3865. 2019. ↩

  3. X. Tang, C. Wang, X. Yuan, and Q. Wang. Non-interactive privacy-preserving truth discovery in crowd sensing applications. In Proceedings of the IEEE INFOCOM. 1988–1996. 2018. ↩

  4. Y. Li, C. Miao, L. Su, J. Gao, Q. Li, B. Ding, Z. Qin, and K. Ren. An efficient two-layer mechanism for privacy-preserving truth discovery. In Proceedings of ACM SIGKDD. ACM, 1705–1714. 2018. ↩

  5. Q. Li, Y. Li, J. Gao, B. Zhao, W. Fan, and J. Han. Resolving conflicts in heterogeneous data by truth discovery and source reliability estimation. In Proceedings of ACM SIGMOD. 1187–1198. 2014. ↩