SEOUL, September 30 (AJP) - A joint team of South Korean and American researchers has won the top paper award from a leading engineering management journal for an artificial intelligence model that reads construction contracts and tells each department on a large project which requirements belong to it.
The team was led by Ko Tae-woo, a professor in Kookmin University's School of Architecture, and Lee Jee-hee, a professor in Chung-Ang University's School of Architecture, working with Texas A&M University professor David Jeong. Their paper won the 2026 Best Paper Award from the Journal of Management in Engineering, Kookmin said Tuesday. The award ceremony was held Monday at Rice University in the United States.
The journal is published by the American Society of Civil Engineers and carries an impact factor of 6, a measure of how often its articles are cited by other researchers. Each year, the American Society of Civil Engineers (ASCE) selects one paper from the journal that it judges strongest in academic and practical contribution, innovation and impact, according to Kookmin.
Large construction projects covering buildings, roads and transport infrastructure bring together many departments and partner organizations. What each of them must do is written into contracts and specifications, the detailed technical documents that set out how the work is to be carried out, and finding one's own duties has meant going through those documents page by page.
The team built an algorithm that uses an AI language model to sort the requirements in complex contract documents by the department responsible for them. It then linked the results to location data from the construction site, creating what it calls an interactive mapping system that marks each department's tasks on a digital map in a different color.
With the system, departments and subcontractors on a road construction site can see on the map where their work is and what they are responsible for, and share related information, without checking lengthy specifications line by line, the university said.
Kookmin said the model recorded high classification accuracy in an empirical study, showing it could be applied at actual construction sites. The university did not release the accuracy figures or say whether the system has been used on an active project.
Ko said being named best paper by a journal of global standing was "a great honor," calling it "a valuable achievement" of a global research team that had worked closely together for a long time.
"I hope this research, which applies AI technology to construction management, will be used as a practical guideline for smart construction that speeds up decision-making on site and reduces the risk of construction delays and work disputes," he said.
[Reference Information]
Journal/Source: Journal of Management in Engineering
Title: Project Requirements Assignment Using BERT-Based Multilabel Classification and Interactive Mapping System
Link/DOI: https://doi.org/10.1061/JMENEA.MEENG-6311
AJP Takeaways
- The Journal of Management in Engineering, published by the American Society of Civil Engineers, gave its 2026 Best Paper Award to a team led by Kookmin University's Ko Tae-woo and Chung-Ang University's Lee Jee-hee, working with Texas A&M University's David Jeong, at a ceremony held Monday at Rice University.
- Ko Tae-woo's team built an AI model that sorts requirements in construction contracts by the department responsible and marks each department's tasks in a different color on a digital map linked to site location data, so teams on a road construction site can find their duties without reading specifications line by line, the university said.
- Kookmin University claimed the model recorded high classification accuracy in an empirical study but, in its announcement Tuesday, did not release the accuracy figures or say whether the system has been used on an active project.
Copyright ⓒ Aju Press All rights reserved.