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.

这项小型企业创新研究（SBIR）I阶段项目的更广泛影响是通过显着增强人工智能（AI）和机器学习（ML）硬件和系统的安全姿势来提高经济和社会福祉的重点，这些姿势越来越无处不在，并且在安全/安全策略策略范围内越来越多地使用。当该项目分析硬件攻击和开拓者的新防御时，它确保了社会依赖医疗保健，金融和国家安全的AI/ML技术的更可靠的基础。商业潜力是巨大的。当开发人员部署这些强化系统时，他们会减轻违反昂贵的违规风险，促进信任并加速采用。经济利益还扩大了与网络攻击有关的支出的减少，并提高了安全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方面的科学知识的前沿。结果，该项目的创新有望加强针对以硬件为中心威胁的关键基础架构，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。小型企业创新研究（SBIR）I阶段I期在解决方案方面，以解决方案的攻击方法，以解决方案，以确保硬件的障碍，以确保硬件/MEL硬件，以确保硬件/MEL Hartging的问题。注射攻击。该项目认识到这些系统对硬件开发的脆弱性，旨在全面分析攻击向量并设计创新的防御机制。拟议的研究旨在采用多层方法，该方法集成了尖端的加密技术和新颖的机器学习算法来增强硬件安全性。通过严格的实验和验证，预期的技术结果包括开发可信赖的硬件模块，建立用于硬件威胁评估的基准测试框架以及创建适应性的，有弹性的防御架构。在AI/ML的背景下，这将大大提高对硬件安全性的科学理解，并有可能为行业实践设定新的标准，同时解决现代计算基础架构中的关键漏洞。该奖项反映了NSF的法定任务，并认为通过使用该基金会的知识分子和更广泛的影响来评估CRITERIA的评估。