A Methodological Approach to Asset Information Management via Knowledge Graphs and Large Language Models

Abstract: Tackling the need of large organizations for a proactive Asset Information Management (AIM) System, a methodological approach to knowledge management applied to built assets portfolios is proposed. It aims at synergically leveraging Knowledge Graphs (KGs) and Artificial Intelligence (AI) technologies to enable analytics on input data. In the theorized pipeline Large Language Models (LLMs) are meant to be used both in the graph creation phase, extracting data from unstructured sources and organizing them according to domain ontologies, as tested on a use-case sample, and in the knowledge extraction phase via queries.