Understanding Fairness in Secure Two-Party and Multi-Party Computation

在安全计算的环境中，一组相互不信任的各方希望计算其输入的某些函数，同时保持安全属性，如隐私、正确性等。一个重要的性质是公平，这大致意味着如果任何人知道计算结果，那么每个人都知道。自20世纪80年代中期以来，人们已经知道，一般来说，如果没有绝对多数诚实的一方，完全公平是不可能的。20多年来，公认的民间传说一直是，没有诚实的多数人，任何小事都不能完全公平地计算出来。PI等人的最新工作表明这个民间传说是错误的，并证明了安全计算中的公平问题是公开的。本项目旨在扩展最初的可行性结果(或证明相应的不可能结果)，并更好地理解密码协议中的公平性。这包括关于部分公平的工作，以及探索作为获得公平的手段的合理性。该项目的教育部分将侧重于研究生教育和与更广泛的研究界的外联努力。

In the setting of secure computation, a set of mutually distrusting parties wish to compute some function of their inputs while preserving security properties such as privacy, correctness, etc. One important property is fairness, which roughly means that if anyone learns the result of the computation, then everyone does.It has been known since the mid-1980s that complete fairness is impossible, in general, without a strict majority of honest parties. For over 20 years, the accepted folklore has been that nothing non-trivial can be computed with complete fairness without an honest majority. Recent work of the PI and others shows that this folklore is false, and demonstrates that the question of fairness in secure computation is wide open.This project seeks to expand the initial feasibility results (or prove corresponding impossibility results) and to develop a better understanding of fairness in cryptographic protocols more generally. This includes work on partial fairness, as well as exploring rationality as a means of obtaining fairness. The educational component of this project will focus on graduate student education and outreach efforts to the broader research community.