SaTC: CORE: Small: Secure Computation with Minimal Interaction

今天，大量的私人数据集被个人和机构存储在个人电子设备和云服务器上。许多具有巨大社会影响的应用程序都可以从对这些数据集进行联合计算的能力中受益。然而，在存在恶意行为者的情况下——无论是执行计算的实体还是外部进程——这种好处都伴随着丢失数据隐私的重大风险。密码学中强大的安全计算范式旨在通过提供一种分布式机制来对个人的私有数据进行计算，从而解决计算与隐私之间的冲突，从而使计算输出之外的任何内容都不会被泄露。安全计算的可行性结果在三十年前就已经确定，未来前景广阔。然而，多年来，安全计算已经发现了有限的实际应用。造成这种差异的主要原因是，已知解决方案的协议复杂性不能满足应用程序的需求。为了使安全计算更接近现实世界的应用，该项目侧重于效率的基本度量——各方之间交互所需的轮数——并旨在完全解决安全计算的轮复杂度，从而产生最佳协议。为了实现这一目标，该项目开发了新的、低开销的方法来实现针对恶意对手的安全性，恶意对手可能会任意偏离其规定的策略。为了促进在现实世界中的自信使用，该项目开发了最小信任设置的解决方案，其中个人不需要信任任何实体或外部过程。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Today, vast amounts of private datasets are stored by individuals and institutions on personal electronic devices and cloud servers. Many applications of tremendous societal impact can benefit from the ability to perform joint computations on these datasets. However, in the presence of malicious actors -- whether it be the entities performing the computation or external processes -- this benefit is associated with a significant risk of losing data privacy. The powerful paradigm of secure computation in cryptography aims to resolve the conflict between computation and privacy by providing a distributed mechanism for computing upon the private data of individuals, so that nothing beyond the output of the computation is revealed.Feasibility results for secure computation were established three decades ago, with an immense promise for the future. Over the years, however, secure computation has found limited real-world applicability. The primary reason for this disparity is that the protocol complexity of known solutions does not meet the demands of applications. To bring secure computation closer to real world applications, this project focuses on a fundamental metric of efficiency -- the number of required rounds of interaction between the parties -- and aims to fully resolve the round complexity of secure computation, leading to optimal protocols. To achieve this goal, the project develops new, low-overhead methods to achieve security against malicious adversaries, who may arbitrarily deviate from their prescribed strategies. To promote confident use in real world, the project develops solutions in the setting of minimal trust, where an individual does not need to trust any entity or external process beyond itself.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.