CAREER: Pushing the Practicality of Secure Multiparty Computation

数据驱动技术的进步通过集中收集用户数据带来了好处。然而，数据集中化和共享也提高了关键的安全和隐私需求。安全多方计算（MPC）允许一组参与方联合计算任何函数，并且只显示输出。这是一个重要的工具，使私人数据计算，而不共享它。本项目的重点是进一步推动MPC协议的实用性。该项目的创新之处在于确定在现实环境中部署MPC协议的障碍，并设计高效，健壮和可扩展的协议来克服这些问题。该项目的广泛意义和重要性有两个方面：1）通过推动MPC的使用，该项目在不牺牲隐私的情况下实现了新的有益应用; 2）该项目涉及教育和外展活动，以加强不同阶段学生的MPC教育，改善社会福利，促进参与的多样性。 该项目侧重于将MPC付诸实践的三个方面。第一个方面是在恶意环境下的安全两方计算协议，其效率和可扩展性与半诚实协议一样。其次，该项目研究了MPC在特定于应用程序的设置中用于大数据时的效率，并改进了基于RAM的MPC协议，用于现实世界的工作负载。第三，该项目开发了新的协议类别，可以将MPC扩展到数百方，并计算大量语句，即使对手可以腐蚀许多参与者。最后，通过学术合作和行业部署，对该奖项的研究成果进行应用和评估。该奖项反映了NSF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。

Advances in data-driven technologies offer benefits through the centralized collection of user data. However, data centralization and sharing also raise critical security and privacy needs. Secure multi-party computation (MPC) allows a set of parties to compute any function jointly and reveals nothing but the output. It is a crucial tool to enable private data computation without sharing it. This project focuses on further pushing the practicality of MPC protocols. The project's novelties are identifying the hurdles of deploying MPC protocols in real-world settings and designing efficient, robust, and scalable protocols to overcome these issues. The project's broader significance and importance are two-fold: 1) by pushing the use of MPC, the project enables new beneficial applications without sacrificing privacy; 2) the project involves educational and outreach activities to enhance MPC education for students at different stages, improve social welfare, and promote diversity of participation. This project focuses on three aspects of bringing MPC to practice. The first aspect is on secure two-party computation protocols in the malicious setting as efficient and scalable as their semi-honest counterparts. Second, the project investigates MPC efficiency when used for big data in application-specific settings and improved RAM-based MPC protocols for real-world workloads. Third, the project develops new classes of protocols that can scale MPC to hundreds of parties and compute huge statements even when the adversary can corrupt many participants. Finally, the research outcomes from this award are applied and evaluated via academic collaboration and industry deployment.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.