Associate 2021-22

Mani Golparvar-Fard

Civil & Environmental Engineering

Golparvar-Fard imageAI-Driven Construction

Professor Golparvar-Fard’s project focuses on preliminary research and planning activities to create foundational ideas, technologies and technical workforces that allow Artificial Intelligence (AI) research to support safe, productive, efficient design, construction and maintenance of national buildings and infrastructure systems. The construction industry is ripe for transformation: inefficient, unsafe, and unpredictable but increasingly digital and open to new technologies. Construction problems expose new challenges for AI, such as learning to solve many related 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. Advances in these areas will form the basis for new approaches to multiple scientific grand-challenges identified by the National Academy of Engineering, such as data-driven construction planning, monitoring work in progress, and real-time worker safety assessment. Professor Golparvar-Fard plans to make preliminary advances to exemplify the opportunities and challenges of AI problems in the construction domain. Specific focus will be on image-based 3D reconstruction of interiors, material recognition, 3D pose estimation, and semi-structured analysis of project schedule text data. The findings in this CAS study will be leveraged to establish the first Institute for AI in Construction. Research outcomes will be disseminated through publications, presentations, and posting of datasets and software.