[51] Making manufacturing knowledge graph more intelligent: A knowledge intelligence management method for manufacturing enterprises

Paper Description extracted from PDF · abstract · 145 words

The rapid development and expanded application of large language models (LLMs) have accelerated the transformation of manufacturing enterprises toward digitalization and intelligence. During this process, knowledge has played an increasingly critical role in supporting enterprises to address new challenges. Knowledge graph (KG) has become one of the core tools for knowledge management in the process of enterprise transformation. Extracting knowledge from multi-source heterogeneous data to construct KG and applying reasoning to new business scenarios is pivotal for intelligent knowledge management in manufacturing enterprises. However, traditional methods for constructing KG, which rely on graph neural networks and deep learning, face challenges such as difficulties in updating knowledge and insufficient practical applications. LLMs have been proven capable of constructing KG from publicly available data. However, when applied to specific domains such as manufacturing, the constructed KG can particularly be prone to issues like incompleteness and low applicability.