PARTNER: Innovating AI for efficient and insightful data transformation

合作伙伴：创新人工智能以实现高效、富有洞察力的数据转换

• 批准号: 2334265
• 负责人: Sudhir Malik
• 金额: $ 279.43万
• 依托单位: University of Puerto Rico Mayaguez
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-09-15 至 2027-08-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2334265&HistoricalAwards=false
• 关键词: PARTNER; Innovating; AI; efficient; insightful

This project is an ExpandAI Partnership between the University of Puerto Rico Mayaguez (UPRM) and the AI Institute for Artificial Intelligence and Fundamental Interactions (IAIFI) In this project, a minority-serving institution leads a new collaboration with an AI Institute focused on scaling up already-established research and education programs at UPRM and to pursue shared, complementary goals around developing AI with use for society in mind and for developing the next generation of AI education and workforce talent. The collaborative research focuses on the development of AI for contributing to breakthroughs in high-energy physics experiments. The project will also build community and new centers of excellence in AI where such activities were not previously well developed. The present era is witnessing scientific discoveries where Artificial Intelligence (AI) plays a decisive role in giving insights into data. The recent explosion of deep learning techniques matched by continuously evolving computing architectures has enabled AI tools to handle higher-dimensional and more complex problems than previously achievable efficiently. The complex, high dimensional data collected by high-energy physics experiments like those at the Large Hadron Collider provide for convergent research in physics and AI. This research is structured around the following AI themes: Optimal Transport in Deep Learning, Physics Informed Uncertainty Quantization, Sparse Dictionary Learning, and Computational aspects of Deep Learning. The intent is to enhance data reduction by transforming the information in the low-level, high-dimensional space into a higher-level and lower-dimensional representation. It will also benefit imaging by enhancing sparse learning, classification, image super-resolution, spectral unmixing, and segmentation. The project emphasizes inclusive participatory design to promote positive societal outcomes of AI innovation. A key goal of the research partnership between UPRM and IAIFI is to significantly increase participation of UPRM in AI research, education, and workforce development. It leverages ongoing research at UPRM on the Compact Muon Solenoid Experiment at the LHC, the Artificial Intelligence Imaging Group’s work on AI research and applications in hyperspectral and biomedical imaging, and IAIFI expertise on advancing foundational AI through novel approaches. The project is partially funded by NSF’s Louis Stokes Alliances for Minority Participation (LSAMP) program within the Division of Equity for Excellence in STEM.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.

该项目是波多黎各马亚圭斯大学（UPRM）和人工智能与基础交互研究所（IAIFI）之间的扩展AI合作伙伴关系。在该项目中，一个少数族裔服务机构与一个人工智能研究所开展了一项新的合作，重点是扩大UPRM已经建立的研究和教育项目，并追求共享，围绕开发人工智能的互补目标，考虑到为社会服务，以及开发下一代人工智能教育和劳动力人才。合作研究的重点是人工智能的发展，为高能物理实验的突破做出贡献。该项目还将建立社区和新的人工智能卓越中心，这些活动以前没有得到很好的发展。当今时代正在见证人工智能（AI）在洞察数据方面发挥决定性作用的科学发现。最近深度学习技术的爆炸式增长与不断发展的计算架构相匹配，使人工智能工具能够比以前更有效地处理更高维度和更复杂的问题。大型强子对撞机（Large Hadron Collider）等高能物理实验收集的复杂、高维数据，为物理学和人工智能的融合研究提供了基础。本研究围绕以下人工智能主题进行：深度学习中的最佳传输、物理不确定性量化、稀疏字典学习和深度学习的计算方面。其目的是通过将低层次、高维空间中的信息转换为更高层次、更低维的表示来增强数据缩减。它还将通过增强稀疏学习、分类、图像超分辨率、光谱解混和分割而有利于成像。该项目强调包容性参与式设计，以促进人工智能创新的积极社会成果。upm和IAIFI之间研究伙伴关系的一个关键目标是显著增加upm在人工智能研究、教育和劳动力发展方面的参与。它利用了upm在大型强子对撞机的紧凑型介子螺线管实验上正在进行的研究，人工智能成像小组在高光谱和生物医学成像方面的人工智能研究和应用工作，以及IAIFI在通过新方法推进基础人工智能方面的专业知识。该项目部分由美国国家科学基金会的路易斯·斯托克斯少数民族参与联盟（LSAMP）项目资助，该项目隶属于STEM卓越平等部。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。