Navigating Standards in Engineering Design through Latent Textual Topology and LLMs

Standards in engineering are crucial for establishing technical specifications and instructions throughout design, production, and application, reflecting standards of best practice. While the volume of engineering standards is crucial for promoting safety, reliability, and efficiency, the large number of standards adversely affects engineers because they have to sift through many standards to find the ones that are most applicable for their work. Moreover, SMEs (Small and Medium-sized Enterprises) are especially impacted, since they often do not have the resources to allocate staff to this task, meaning that most employees have to deal with this in addition to their primary duties. The aim of this research is to address this problem

by developing a model for using language embedding and Large Language Models (LLMs) to assist in the navigation of engineering standards. The study attempts to answer the question of whether LLMs (Large Language Models) – specifically BERT (Bidirectional Encoder Representations from Transformers) – can be used to uncover hidden spatial relationships between texts within engineering standards. The hope is that this approach will enable the automation of more sophisticated recommendations and organizational structures for designers, resulting in less time spent discovering and applying standards, thus improving the overall quality of projects and compliance with best practices.

Message: Hi FE, I need help with “Navigating Standards in Engineering Design through Latent Textual Topology and LLMs” in “Computer Science & Engineering”