[论文解读] A Blockchain-Based Approach for Saving and Tracking Differential-Privacy Cost

Yang Zhao, Jun Zhao+4 more|arXiv (Cornell University)|Jan 25, 2020

Privacy-Preserving Technologies in Data参考文献 50被引用 5

一句话总结

本文提出了一种基于区块链的系统，通过在重复查询中重用带噪声的查询响应，追踪并最小化差分隐私（DP）成本。通过动态调整先前噪声输出的重用比例，并将所有查询和隐私支出记录在防篡改的区块链上，该系统确保了透明性并防止隐私预算被过度使用，显著降低了累积的DP成本。

ABSTRACT

An increasing amount of users' sensitive information is now being collected for analytics purposes. To protect users' privacy, differential privacy has been widely studied in the literature. Specifically, a differentially private algorithm adds noise to the true answer of a query to generate a noisy response. As a result, the information about the dataset leaked by the noisy output is bounded by the privacy parameter. Oftentimes, a dataset needs to be used for answering multiple queries (e.g., for multiple analytics tasks), so the level of privacy protection may degrade as more queries are answered. Thus, it is crucial to keep track of the privacy spending which should not exceed the given privacy budget. Moreover, if a query has been answered before and is asked again on the same dataset, we may reuse the previous noisy response for the current query to save the privacy cost. In view of the above, we design and implement a blockchain-based system for tracking and saving differential-privacy cost. Blockchain provides a distributed immutable ledger that records each query's type, the noisy response used to answer each query, the associated noise level added to the true query result, and the remaining privacy budget in our system. Furthermore, since the blockchain records the noisy response used to answer each query, we also design an algorithm to reuse previous noisy response if the same query is asked repeatedly. Specifically, considering that different requests of the same query may have different privacy requirements, our algorithm (via a rigorous proof) is able to set the optimal reuse fraction of the old noisy response and add new noise (if necessary) to minimize the accumulated privacy cost. Experimental results show that the proposed algorithm can reduce the privacy cost significantly without compromising data accuracy.

研究动机与目标

• 解决在数据集被多次查询时，差分隐私（DP）预算未被追踪且被过度消耗的挑战。
• 减少对同一数据集重复查询所导致的累积隐私成本。
• 利用区块链技术提供透明、防篡改的DP预算使用审计记录。
• 实现动态、最优的先前计算噪声响应的重用，以最小化新增噪声的添加并保护隐私预算。
• 通过区块链上的智能合约确保一旦预算耗尽，将不再产生任何额外的隐私成本。

提出的方法

• 设计一种新颖算法，计算重复查询中先前噪声响应的最优重用比例，以最小化总隐私成本。
• 集成高斯机制，为查询结果添加噪声，其中敏感度值预先定义以提高效率。
• 实施基于区块链的系统，将所有查询、响应、噪声水平和剩余隐私预算记录为不可篡改的交易。
• 在区块链上使用智能合约强制执行预算耗尽——一旦隐私预算耗尽，自动阻止进一步查询。
• 在采用权威证明（Proof-of-Authority）共识的私有区块链上部署系统，以保护查询和智能合约的隐私。
• 提供预定义的标准查询列表，以限制查询的多样性，防止敏感查询内容泄露。

实验结果

研究问题

• RQ1 当同一查询在数据集上被多次发出时，如何最小化差分隐私成本？
• RQ2 如何构建一个防篡改、透明的系统，以追踪并强制执行差分隐私预算限制？
• RQ3 重用噪声查询响应的最优策略是什么，以减少累积隐私成本？
• RQ4 如何利用区块链技术确保问责性并防止隐私预算过度使用？
• RQ5 在查询重用、噪声添加和隐私预算保护之间，实际的权衡是什么？

主要发现

• 所提出的算法通过基于查询相似性和隐私参数智能重用先前噪声响应，显著降低了累积隐私成本。
• 基于区块链的系统提供了端到端的透明性，使数据所有者和用户能够审计所有查询活动，并验证隐私预算未被超出。
• 区块链上的智能合约强制执行自动预算耗尽，确保一旦达到限额，将不再产生任何额外的隐私成本。
• 通过预定义查询类型和敏感度，该系统支持在实际部署中高效、可扩展地管理差分隐私。
• 通过私有区块链和访问控制实现安全私密的部署，降低了智能合约和查询泄露的风险。
• 该方法在数学上被证明能最小化重复查询的总隐私成本，并对重用策略进行了严格的理论验证。