MRI: Acquisition of Dolly Sods GPU Cluster for Accelerated High-Performance Computing and Applications in Machine Learning and Artificial Intelligence in West Virginia

MRI：收购 Dolly Sods GPU 集群，以加速西弗吉尼亚州机器学习和人工智能的高性能计算和应用

• 批准号: 2117575
• 负责人: Aldo Romero
• 金额: $ 109.94万
• 依托单位: West Virginia University Research Corporation
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-10-01 至 2024-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2117575&HistoricalAwards=false
• 关键词: MRI; Acquisition; Dolly; Sods; GPU

This project will enable West Virginia University (WVU) to acquire a special-purpose graphics processing unit (GPU) cluster called Dolly Sods. Dolly Sods will be a critical driver of WVU's goal of developing capabilities in utilizing big data, artificial intelligence (AI), and machine learning (ML) to enable transformational research in a broad range of fields encompassing drug development, interstellar phenomena, biometrics, material design, and business logistics and management. In conjunction with other institutions of higher learning in West Virginia, development of these capabilities will in turn lead to training an AI-ready West Virginian workforce that can leverage their skills to strengthen current relationships with regional high performance computing (HPC) centers as well as forging new relationships with federal and industrial partners. The creation of training opportunities for first generation college students, female students, and those from marginalized communities will aid in the diversification of the computationally intensive workforce and will be invaluable to West Virginia. Finally, this acquisition will contribute to keeping the United States at the forefront of AI development.The rapid adoption of hardware accelerators (e.g., GPUs) in the research computing community has facilitated massively increased compute capability and application of ML and AI approaches to solving big data problems in every STEM field and non-traditional fields like business data analytics. The requested acquisition of the Dolly Sods cluster will enable cutting-edge research for efforts in diagnostic imaging of tumors, high-throughput screening of small molecule drug design, on-the-fly detection of interstellar phenomena, design and optimization of data compression algorithms used in space flight, computer vision of medical images, and informatics of business-based managerial decisions and system/process analysis, among others. This research will put WVU at the forefront of the movement to integrate ML and AI into the fabric of data-driven research, allowing for timely and relevant scientific and societal issues to be addressed as well as providing training opportunities for the next generation of data scientists, led by 24 faculty in conjunction with 92 postdocs, graduate students, and undergraduate students. Dolly Sods will also serve as an integral teaching resource for a newly formed undergraduate degree in Data Science, which will incorporate ML and AI throughout the curriculum. Combined with outreach to statewide institutions of higher learning, Dolly Sods will drive continued efforts in training a diverse HPC workforce, drawing from communities historically underrepresented in STEM, including first generation college students, women, and other underrepresented groups. The project will strengthen partnerships with regional federal labs (DOE National Energy Technology Laboratory, NOAA Environmental Security Computing Center, NASA Independent Validation and Verification Facility, and the FBI Criminal Justice Services Division), industrial partners (Leidos and PPG), and the nation’s leading center for ML and AI (Pittsburgh Supercomputing Center (PSC)). It will further WVU’s long-term goal of transforming the economically disadvantaged region of Appalachia into a top-level destination for investment from the technology sector.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.

该项目将使西弗吉尼亚大学（WVU）获得一个名为Dolly Sods的专用图形处理单元（GPU）集群。Dolly Sods将成为WVU发展利用大数据，人工智能（AI）和机器学习（ML）的能力的目标的关键驱动力，以实现广泛领域的转型研究，包括药物开发，星际现象，生物识别，材料设计以及商业物流和管理。与西弗吉尼亚州的其他高等院校一起，这些能力的发展将反过来导致培训一个人工智能就绪的西弗吉尼亚州劳动力，可以利用他们的技能来加强与区域高性能计算（HPC）中心的现有关系，并与联邦和工业合作伙伴建立新的关系。为第一代大学生、女学生和来自边缘化社区的学生创造培训机会将有助于计算密集型劳动力的多元化，对西弗吉尼亚州来说将是无价的。最后，这次收购将有助于保持美国在人工智能发展的前沿。硬件加速器的快速采用（例如，研究计算社区中的GPU（图形处理器）促进了计算能力的大幅提高，以及ML和AI方法的应用，以解决每个STEM领域和非传统领域（如商业数据分析）的大数据问题。请求收购Dolly Sods集群将使尖端研究能够在肿瘤诊断成像，小分子药物设计的高通量筛选，星际现象的飞行检测，太空飞行中使用的数据压缩算法的设计和优化，医学图像的计算机视觉，以及基于业务的管理决策和系统/流程分析的信息学等方面做出努力。这项研究将使WVU处于将ML和AI整合到数据驱动研究结构中的运动的最前沿，从而能够及时解决相关的科学和社会问题，并为下一代数据科学家提供培训机会，由24名教师与92名博士后，研究生和本科生共同领导。Dolly Sods还将作为新成立的数据科学本科学位的完整教学资源，该学位将在整个课程中纳入ML和AI。结合对全州高等院校的推广，Dolly Sods将继续努力培训多元化的HPC员工队伍，从STEM历史上代表性不足的社区中汲取经验，包括第一代大学生，女性和其他代表性不足的群体。该项目将加强与区域联邦实验室（DOE国家能源技术实验室、NOAA环境安全计算中心、NASA独立验证和核查设施以及FBI刑事司法服务部）、工业合作伙伴（Leidos和PPG）的合作伙伴关系。）以及美国领先的机器学习和人工智能中心（匹兹堡超级计算中心（PSC））。它将进一步西弗吉尼亚大学的长期目标，改造的经济落后地区的阿巴拉契亚成为一个顶级的投资目的地，从技术部门。这个奖项反映了NSF的法定使命，并已被认为是值得通过评估使用基金会的智力价值和更广泛的影响审查标准的支持。