CAREER: Unified Reference-Free Early Detection of Hardware Trojans via Knowledge Graph Embeddings

职业：通过知识图嵌入对硬件木马进行统一的无参考早期检测

• 批准号: 2238976
• 负责人: Inna Partin-Vaisband
• 金额: $ 60万
• 依托单位: University of Illinois at Chicago
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-08-01 至 2028-07-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2238976&HistoricalAwards=false
• 关键词: CAREER; Unified; Reference; Free; Early

In recent years, cloud computing has risen to unparalleled societal importance, as it enables a broad range of applications, including healthcare services, social media, banking and electronic commerce. As a result, cloud services have access to the digital data of millions of users and become a lucrative target for criminals and cyberattacks. Sensitive data are stolen from cloud servers, incurring significant financial losses to those impacted. To address the shortcomings of current protection technologies in cloud computing, this project employs advanced cryptography and introduces new privacy protections that allow data to remain secure even when the cloud service is compromised. The project's novelties are in the judicious design of computers that can understand and process encrypted information such that, even if data are stolen or under attack, they will be incomprehensible to the attacker. The project's broader significance and importance is in addressing the urgent need of the cloud computing industry to mitigate the impact of cyberattacks and to protect the privacy of users, which is aligned with the mission of the National Science Foundation. The key observation in this project is that data are immune to leakage threats as long as they remain encrypted at all times, including during processing. Towards that end, the investigator employs a special form of encryption, called Fully Homomorphic Encryption, to protect general-purpose computation using diverse hardware platforms such as graphics processing units. Specifically, this project defines imperative models of computation that improve the programmability of encrypted data processing, without sacrificing execution performance. These models are complemented with real-life applications such as privacy-preserving machine learning and data mining that harness the power of encrypted computation. The results of this project are integrated with educational activities, including web-based learning environments and cybersecurity competitions.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.

近年来，云计算已经上升到前所未有的社会重要性，因为它支持广泛的应用，包括医疗服务，社交媒体，银行和电子商务。因此，云服务可以访问数百万用户的数字数据，并成为犯罪分子和网络攻击的有利可图的目标。敏感数据从云服务器中被盗，给受影响的人带来重大经济损失。为了解决云计算中当前保护技术的缺点，该项目采用了先进的加密技术，并引入了新的隐私保护措施，即使在云服务受到损害时，数据也能保持安全。该项目的新颖之处在于计算机的明智设计，这些计算机可以理解和处理加密信息，即使数据被盗或受到攻击，攻击者也无法理解。该项目更广泛的意义和重要性在于解决云计算行业减轻网络攻击影响和保护用户隐私的迫切需求，这与美国国家科学基金会的使命是一致的。该项目中的关键观察结果是，只要数据始终保持加密，包括在处理过程中，数据就不会受到泄漏威胁。为此，研究人员采用了一种特殊的加密形式，称为全同态加密，以保护使用各种硬件平台（如图形处理单元）的通用计算。具体来说，该项目定义了命令式计算模型，可以提高加密数据处理的可编程性，而不会牺牲执行性能。这些模型与现实生活中的应用程序相补充，例如保护隐私的机器学习和利用加密计算能力的数据挖掘。该项目的成果与教育活动相结合，包括基于网络的学习环境和网络安全竞赛。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。