SBIR Phase I: SaiFE: Trusted AI with Hardware Security Enforcement

SBIR 第一阶段：SaiFE：具有硬件安全实施的可信人工智能

• 批准号: 2333126
• 负责人: Emre Karabulut
• 金额: $ 27.28万
• 依托单位: MITHRILAI CORP.
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2024
• 资助国家: 美国
• 起止时间: 2024-02-15 至 2025-01-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2333126&HistoricalAwards=false
• 关键词: SBIR; Phase; SaiFE; Trusted; AI

The broader impact of this Small Business Innovation Research (SBIR) Phase I project is centered on elevating economic and societal well-being by significantly enhancing the security posture of Artificial Intelligence (AI) and Machine Learning (ML) hardware and systems, which are increasingly ubiquitous and used in safety/security-critical applications. As this project analyzes hardware attacks and pioneers new defenses, it ensures a more reliable foundation for AI/ML technologies that society relies upon for healthcare, finance, and national security. The commercial potential is substantial; as developers deploy these fortified systems, they mitigate the risk of costly breaches, fostering trust and accelerating adoption. Economic benefits also extend to a reduction in expenditure related to cyberattacks and an increase in market competitiveness for secure AI/ML products. Furthermore, by deepening understanding of hardware vulnerabilities and defense mechanisms, the project pushes the frontiers of scientific knowledge in cybersecurity. As a result, the innovations from this project are poised to reinforce critical infrastructure against hardware-centric threats, thereby safeguarding the digital economy and reinforcing the United States' leadership in secure technological advancements.This Small Business Innovation Research (SBIR) Phase I project conducts a transformative approach to addressing the acute problem of securing AI/ML hardware systems against emerging hardware attacks such as side-channel and fault injection attacks. Recognizing the vulnerability of these systems to hardware exploitation, the project aims to comprehensively analyze the attack vectors and devise innovative defense mechanisms. The proposed research is set to employ a multi-layered methodology that integrates cutting-edge cryptographic techniques and novel machine-learning algorithms to enhance hardware security. Through rigorous experimentation and validation, the anticipated technical results include the development of trusted hardware modules, the establishment of a benchmarking framework for hardware threat assessment, and the creation of adaptable, resilient defense architectures. This will significantly advance scientific understanding of hardware security in the context of AI/ML, potentially setting a new standard for industry practices, while addressing a critical vulnerability in modern computing infrastructure.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.

小型企业创新研究(SBIR)第一阶段项目的更广泛影响集中在通过显著增强人工智能(AI)和机器学习(ML)硬件和系统的安全状态来提高经济和社会福祉，这些硬件和系统越来越普遍，并在安全/安全关键应用中使用。随着该项目分析硬件攻击并开创新的防御措施，它确保了社会在医疗、金融和国家安全方面所依赖的AI/ML技术有更可靠的基础。商业潜力是巨大的；随着开发人员部署这些加固的系统，他们降低了代价高昂的入侵风险，促进了信任并加快了采用。经济效益还包括减少与网络攻击有关的支出，以及提高安全AI/ML产品的市场竞争力。此外，通过加深对硬件漏洞和防御机制的理解，该项目推动了网络安全科学知识的前沿。因此，该项目的创新将加强关键基础设施以抵御以硬件为中心的威胁，从而保护数字经济并加强美国在安全技术进步方面的领导地位。该小型企业创新研究(SBIR)第一阶段项目采用了一种变革性的方法来解决保护AI/ML硬件系统免受新出现的硬件攻击(如侧通道和故障注入攻击)的严重问题。认识到这些系统对硬件攻击的脆弱性，该项目旨在全面分析攻击载体并设计创新的防御机制。这项拟议的研究将采用一种多层次的方法，将尖端的密码技术和新颖的机器学习算法相结合，以增强硬件安全。通过严格的实验和验证，预期的技术成果包括开发受信任的硬件模块，建立硬件威胁评估的基准框架，以及创建适应性强、弹性强的防御架构。这将显著提高对AI/ML背景下硬件安全的科学理解，潜在地为行业实践设定新的标准，同时解决现代计算基础设施中的一个关键漏洞。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。