Treffer: Intelligent Processing of Design Notices in Engineering Procurement Construction Projects.

The accumulation and delayed processing of notices generated during the engineering construction process have a significant impact on project settlement and, thus, project cost. Currently, there is a lack of research on intelligent notice processing. Although large language models (LLMs), such as ChatGPT, have demonstrated exceptional performance in natural language processing, their effectiveness in specific vertical fields, such as construction engineering, is limited due to a lack of specialized training. In light of this, this study proposes a knowledge-augmented language model for intelligently processing design notices in EPC (engineering–procurement–construction) projects. This method consists of the following three key components: database construction, price retrieval, and prompt development. During database construction, exception detection was introduced to ensure data quality, and an appropriate database framework was proposed. The price retrieval module features innovative retrieval rules for improved efficiency and accuracy. Prompt development was based on mainstream methods, which were tailored for this task. The result of processing notices includes cost analysis and claimability judgement. The method achieved promising results in experiments with real project data. Based on these results, the paper discusses the model's advantages, application scenarios, and input text requirements, providing insights and suggestions for future research. [ABSTRACT FROM AUTHOR]

Copyright of Buildings (2075-5309) is the property of MDPI and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)