Unlocking ancient wisdom with modern tools: A new approach to the revitalization of ancient texts based on generative artificial intelligence

In the fields of library science and information science, the gap between the obscure language of ancient texts and the cognitive abilities of the average reader is the primary obstacle to the widespread dissemination of the rich wisdom in traditional cultural heritage. The rapid development of generative artificial intelligence technology has provided the conditions for researchers to transition from sequential organization of bibliographic references to the deconstruction and reorganization of knowledge elements. This study, reveals the intrinsic relationship patterns between multi-granularity knowledge in ancient texts from perspectives such as logical relationships, frame semantic associations, hierarchical structures, and feature associations, and constructs a multidimensional knowledge representation model for ancient texts. Furthermore, a YAML prompt template was designed by integrating large language models with the deconstruction of multi-granularity knowledge elements from ancient texts, and experiments were conducted using the Chinese classic “Records of the Grand Historian: Biographies.” The results indicate that the method developed in this study for the deconstruction and reorganization of knowledge elements in ancient texts exhibits notable characteristics such as multi-granularity, cross-document application, refinement, multiple perspectives, efficiency improvement, and scalability, demonstrating potential for application in the field of knowledge services for ancient texts.