Theory and Practice of Secure Computation

在我们为有用的目的收集数据的愿望和保护个人隐私的愿望之间存在着根本的紧张关系。典型的例子是投票，在投票中，我们希望在不显示任何个人选票的情况下将选票相加。安全计算是密码学的一个分支，它调解并在某些情况下转换这种紧张，使我们能够满足这些相互冲突的需求，达到比天真看起来可能的水平要高得多的水平。这项研究主要涉及安全计算的基本理论，确定在什么条件下可以从多个数据来源获取有用信息，同时最大限度地保护这些数据的隐私。它还考虑了安全计算可能特别有效和实用的特殊情况。实用的安全计算解决方案将对许多应用有用，包括电子投票、电子商务和访问控制系统。这项研究的具体目标包括：对一般类别的安全计算系统的能力进行分类，特别是那些具有概率输出行为的安全计算系统。：进一步发展安全计算基元之间的无差错减少理论，其中目标是以确定性而不是仅仅以高概率实现的。：为具有复杂的计票机制(例如，即时决选投票)的投票系统开发更有效的协议，以及其他偏好聚合问题。：找到方法将现有的信息服务用作实现安全计算的手段，以及作为弥合抽象协议和真实系统之间差距的资源。努力程度声明我们注意到，建议首席研究人员的夏季支持量已减少到每年一个月。在建议的支助水平下，首席调查员将尽一切努力满足项目的最初范围以及他的努力水平。

There is a fundamental tension between our desire to aggregate data for useful purposes and our desire to protect the privacy of individuals. The canonical example is voting, in which we wish to add up votes without revealing any individual votes. Secure computation is a branch of cryptography that mediates and in some cases transforms this tension, allowing us to satisfy these conflicting needs to a far greater level than naively would seem possible. The research is primarily concerned with the basic theory of secure computation, determining under what conditions it is possible to obtain useful information from multiple sources of data while maximally protecting the privacy of this data. It also considers special cases under which secure computation may be particularly efficient and practical. Practical secure computation solutions would be useful for a number of applications, including electronic voting, e-commerce and access control systems.Specific goals of this investigation include:: Classifying the power of general classes of secure computation systems, particularly those with probabilistic output behavior.: Further developing the theory of errorless reductions among secure computation primitives, in which the goals are achieved with certainty as opposed to merely with high probability.: Developing more efficient protocols for voting systems with sophisticated vote-counting mechanisms (e.g., instant runoff voting), and for other preference aggregation problems.: Finding ways to use extant information services as a means for implementing secure computations and as a resource for bridging the gap between abstract protocols and real systems.Level of effort statementWe note that the amount of summer support recommended for the principal investigator has been reduced to one month per year. At the recommended level of support, the principal investigator will make every attempt to meet the original scope of the project, as well as his level of effort.1