AI Institute: Planning: Construction

点击翻译按钮获取中文摘要

Improvements in the design, construction and operation of buildings and infrastructure projects will have a tremendous impact on the competitiveness of the U.S. Construction industry, by making projects more efficient, cheaper, faster, safer and more equitable.   Nevertheless, enhancing project workflows poses problems that challenge current understanding of Artificial Intelligence (AI).  This National Artificial Intelligence Research Institutes Planning award supports research and coordination activities to build collaborations among AI researchers, construction researchers, and industry partners, with the aim of forming an Institute for AI in Construction. Key goals are the identification of critical scientific problems, performance metrics, data sources, and future grand challenges.  The team will also develop opportunities for educational activities that attract and retain skilled workforce in the industry. To do so, the team will offer a new course on Construction AI, formulate a new "CS+Construction" major, and design a Masters' capstone project sponsorship program, each with a strong element of entrepreneurship education. A mentoring program that aids students from underrepresented groups will be established particularly around their engagement in construction projects in the form of Women and Minority-Owned Business Enterprise (WMBE) firms. Many planning, monitoring, and control workflows involved in the design, construction, and operation of the built environment expose new challenges and opportunities for research in computer vision, natural language processing, and machine learning.   Key technical problems, as identified by the National Academy of Engineering, include data-driven construction planning, monitoring work in progress, and real-time worker safety assessment.  Solving these problems requires fundamental research in AI, such as: machine learning with many interconnected small-data problems; optimizing for application-specific objectives; leveraging both recognition and correspondence to recover geometry from images; and learning from loosely structured text documents.  This research will identify AI problems in the construction domain that can serve as model problems, uncover novel conceptual challenges to AI research from construction applications, and identify likely dataset needs to support future research on AI in construction.  Research findings will be disseminated through publications, presentations, and posting of datasets and software through a central project website.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.