SCIPE: Chishiki.ai: A sustainable, diverse, and integrated CIP community for Artificial Intelligence in Civil and Environmental Engineering

Chishiki.ai是一个涵盖人工智能(AI)和土木工程与环境工程(CEE)实践的CI专业人士(CI)的综合社区，以支持美国的基础设施，与2020年国家人工智能倡议法案和2022年基础设施投资和就业法案确定的主要领域保持一致。Chishiki项目采用四项战略，通过(1)通过研究峰会、研究生和本科生奖学金、联合研究计划和产业伙伴关系等举措，促进CIP和领域专家之间的合作；(2)提供由AI提供动力的个性化和可扩展的学习环境；(3)开发创新的AI使能的CI架构，以实现可重复和高效的工作流程；以及(4)通过招聘和研究计划与历史上代表性不足的机构接触，创建一个多样化、可持续的CIP社区。Chishiki提供同行指导支持，并与NSF Access计算科学支持网络(CSSN)合作，在与CI和CEE相关的研究活动中支持CI专业人员。通过积极的实践社区，将CIPS整合到中东欧研究中，为专业发展、合作和福祉提供机会。该项目将公布在中东欧建立伙伴关系、扩大采用和普及传播与信息解决方案的最佳做法。Chishiki提供人工智能增强的CI解决方案，并支持一个整合和多样化的CIP社区，致力于改造土木工程和环境工程。通过Chishiki.ai，该项目为CI专业人员开发新课程，以构建和支持复杂的CI框架，以促进人工智能驱动的研究创新。关于AI4CI和CI4AI的课程涵盖了支持AI的编程、高性能计算(HPC)系统的AI增强性能调整、AI驱动的知识发现和管理，以及构建大规模可用于生产的AI系统。关于科学机器学习的课程探索了可解释人工智能、可微编程和不确定性传播的技术，从而使CI专业人员能够理解在CEE中人工智能的需求和使用。Chishiki项目开发了一个新颖的、可扩展的学习环境，它通过强化学习来构建上下文感知的大型语言模型，以生成个性化的测验和解释。个性化的AI导师有助于生成个性化的测验和定制的解释，以满足个人的需求和学习能力。这些可扩展和个性化的AI导师支持的课程将在康奈尔虚拟研讨会(CVW)学习平台上以开放访问内容的形式提供，接触到广泛的CIPS社区。为了加速人工智能增强的研究，该项目支持开发基于图形神经网络和用于优化和工程设计的可微模拟的复杂人工智能代理，开发框架以在内存有限的边缘设备上部署基本的人工智能模型，用于结构健康监测和交通规划，以及高性能计算系统，以开发支持人工智能的CEE的示范应用。Chishiki项目还支持人工智能辅助代码开发，以加快科学研究。该项目的交付成果将在现有的NSF资助的平台DesignSafe和德克萨斯高级计算中心(TACC)上提供，从而扩大人工智能增强型CI创新的采用和集成。这些发展将作为公开课程内容和开放源码解决方案向公众开放，以便更广泛地传播。该项目的目标是通过这个个性化和可扩展的学习平台，使全国500多家CIPS受益，并在全球培训30多万用户。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Chishiki.ai is an integrated community of CI professionals (CIPs) across artificial intelligence (AI) and civil and environmental engineering (CEE) practices to bolster U.S. infrastructure, aligning with thrust areas identified by the 2020 National Artificial Intelligence Initiative Act and the 2022 Infrastructure Investment and Jobs Act. THe Chishiki project adopts four strategies to build a sustainable, diverse, and integrated community of CI professionals for AI in CEE by (1) fostering collaboration between CIPs and domain experts through initiatives such as research summits, graduate and undergraduate fellowships, joint research initiatives, and industrial partnerships; (2) offering personalized and scalable learning environments powered by AI; (3) developing innovative AI-enabled CI architectures for reproducible and efficient workflows; and (4) creating a diverse, sustainable CIP community through engagement with historically underrepresented institutions through recruitment and research initiatives. Chishiki offers peer mentoring support and works with the NSF ACCESS Computational Science Support Network (CSSN) to support CI professionals in research activities related to CI and CEE. The integration of CIPs into CEE research is enabled through an active community of practice, providing opportunities for professional development, collaboration, and well-being. The project will publish best practices on partnerships, broadening adoption, and democratizing access to CI solutions in CEE. Chishiki offers AI-enhanced CI solutions and supports an integrated and diverse CIP community dedicated to transforming Civil and Environmental Engineering.Through Chishiki.ai, the project develops new courses for CI professionals to build and support sophisticated CI frameworks that foster AI-driven research innovations. The courses on AI4CI and CI4AI cover AI-enabled programming, AI-enhanced performance tuning of High-Performance Computing (HPC) systems, AI-driven knowledge discovery and curation, and building large-scale production-ready AI systems. The course on Scientific Machine Learning explores techniques for explainable AI, differentiable programming, and uncertainty propagation, thus enabling CI professionals to understand the need and use of AI in CEE. The Chishiki project develops a novel, scalable learning environment by building context-aware Large Language Models through reinforcement learning to generate personalized quizzes and explanations. The personalized AI tutor facilitates generating individualized quizzes and customized explanations to suit the individual's needs and learning abilities. The scalable and personalized AI tutor-powered courses will be available as open-access content on the Cornell Virtual Workshop (CVW) learning platform, reaching a broad community of CIPs. To accelerate AI-enhanced research, the project supports the development of sophisticated AI surrogates based on graph neural networks and differentiable simulations for optimization and engineering design, develops frameworks to deploy foundational AI models on memory-limited edge devices for structural health monitoring and transportation planning, and HPC systems to develop exemplar applications of AI-enabled CEE. The Chishiki project also supports AI-assisted code development to accelerate scientific research. The project's deliverables will be available on existing NSF-funded platforms, DesignSafe and the Texas Advanced Computing Center (TACC), broadening the adoption and integration of AI-enhanced CI innovations. The developments will be publicly accessible as open-course content and open-source solutions for broader dissemination. The project goal is to benefit more than 500 CIPs nationwide and to train more than 300,000 users worldwide through this personalized and scalable learning platform.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.