SHF: Small: Fault Model Evaluation and Discovery

集成电子电路（ic）今天无处不在，因为它们影响着我们日常生活的方方面面，从各种形式的移动设备（智能手机、笔记本电脑等）到各种交通方式（汽车、飞机等）。电子技术的进步也为最近和即将到来的机器学习、人工智能、自动驾驶汽车等领域的进步奠定了基础。该项目的研究工作将通过帮助领先公司确保电子产品在长时间内正常可靠地工作，从而有助于电子半导体行业的持续发展。此外，进行的研究将包括本科生研究人员，研究结果将传播给研究界，并纳入高等教育课程。这项研究将专门研究如何对电子故障进行建模，对于发现的任何缺陷，将开发更准确的新模型。众所周知，模型的有效性很大程度上取决于集成电路及其制造，但不知道哪些模型将被证明是无效的。此外，特别是对于尖端技术，预计还需要为最先进的集成电路提供新的，更准确的故障模型。因此，本研究的目标是开发一种全面的数据驱动方法，用于断层模型的持续评估和开发。将分析来自故障ic的数据，以了解故障的可能性和类型。对于现有模型与实测数据之间的预期不匹配，将提出并评估新模型。成功实现这项研究的目标将使集成电路制造商能够以更低的成本生产更可靠的集成电路。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。

Integrated electronic circuits (ICs) today are ubiquitous in that they affect every aspect of our daily lives, ranging from use within all forms of mobile devices (smart phones, laptops, etc.) to all modes of transportation (cars, aircraft, etc.). Advances in electronics also form the foundation for the recent and forthcoming advances in machine learning, artificial intelligence, autonomous vehicles, etc. The research work stemming from the project will aid the continued advancement of the electronic semiconductor industry by helping leading companies ensure that electronics work properly and reliably over long time periods. Additionally, the research conducted will include undergraduate student researchers, and the results will be disseminated to the research community, and incorporated into post-secondary curriculum.The research will specifically examine how electronic failures are modeled, and for any shortcomings discovered, new models that are more accurate will be developed. While it is widely known that model effectiveness greatly depends on the IC and its fabrication, it is not known which models will be proven ineffective beforehand. In addition, especially for cutting-edge technologies, it is expected that there also will be a need for new, more accurate fault models for state-of-the-art ICs. Therefore, the objective of this research is to develop a comprehensive data-driven methodology for the continuous evaluation and development of fault models. Data from failed ICs will be analyzed to understand the likelihood and types of failures. For the expected mismatch between existing models and measured data, new models will be proposed and evaluated. Successfully meeting the objective of this research would will enable IC manufacturers to produce more reliable ICs at much lower cost.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.