Center: IUCRC Phase II Rutgers University: Center for Accelerated and Real Time Analytics (CARTA)

加速和实时分析是智能数据革命的前沿，一方面是由互联网连接的传感器硬件的进步推动的，另一方面是加速的AI/ML分析。这些技术包括人工智能/机器学习、认知、量子、神经形态和边缘计算，结合了新型的感官和可编程硬件组件、主动存储、闪存、内存云系统、嵌入式传感器等。它们涉及国防和网络安全、天气和野火预测、健康和疾病监测、工业物联网等诸多领域。在IUCRC CARTA第二阶段，我们将开发新的加速和实时的方法来处理与我们的工业和机构伙伴相关的这些技术。我们将利用我们在人工智能/机器学习和可编程网络系统方面的前沿研究来开展合作的CARTA研究。我们的研究领域包括加速医学图像分析，用于医学和视觉/语言模型的可解释/可解释的AI/ML，用于加速动力系统的数字孪生，以及用于改进理解的基于物理的AI/ML建模，用于加速分析的可编程耦合软件和硬件。我们的研究将主要是“应用中立”的。因此，我们将开发的技术可以应用于多个行业领域，包括网络安全、医疗保健、动态系统、环境和商业智能。我们将探索的技术包括半/无监督学习、联邦学习、可编程架构、语言/视觉模型、实时仿真和区块链。我们的项目将为企业提供更多的高技能劳动力。我们将与工业界广泛传播这项合作研究的结果，在适当的期刊和会议上发表。我们还建议与其他想要复制这一方法的非carta机构分享我们的研究结果。我们将利用我们针对代表性不足的本科生/研究生的项目，招募他们到我们的行业合作伙伴那里进行暑期实习，并参与CARTA项目的研究。数据科学硕士课程将用于学生完成CAPSTONE项目。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。

Accelerated and real-time analytics is a leading edge of the smart data revolution, pushed by advances in internet-connected sensor hardware on one side and accelerated AI/ML analytics on the other. These technologies include AI/ML, cognitive, quantum, neuromorphic, and edge computing combined with new classes of sensory and programmable hardware components, active storage, flash storage, in-memory cloud systems, embedded sensors, etc. They address many pursuits of national defense and cybersecurity, weather and wildfire prediction, health- and disease-monitoring, Industrial IoT, etc. In IUCRC CARTA Phase II, we will develop new accelerated and real-time approaches to these technologies relevant to our industrial and agency partners. We will leverage our cutting-edge research in AI/ML and programmable cyber-systems to conduct collaborative CARTA research. Our research areas include accelerated medical image analytics, explainable/interpretable AI/ML for medical and vision/language models, digital twins for dynamical systems with acceleration, and physics-based modeling with AI/ML for improved understanding, programmable coupled software and hardware for accelerated analytics. Our research will be primarily “application neutral.” Therefore, the techniques we will develop can be applied with modifications across multiple industry sectors, including cyber security, healthcare and medical, dynamic systems, environment, and business intelligence. The technologies we will explore include semi/un-supervised learning, federated learning, programmable architectures, language/vision models, real-time simulations, and blockchains. Our project will benefit businesses by providing a larger pool of highly skilled workforce. We will disseminate the results of this collaborative research with industry widely by publishing them in appropriate journals and conferences. We also propose to share the results of our research with other non-CARTA institutions that want to duplicate this approach. We will leverage our programs for underrepresented undergraduate/graduate students to recruit them for summer internships at our industry partners and research in CARTA projects. The MS in Data-Science Program will be used for students to do CAPSTONE projects.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.