Unlocking Relations in Classical Chinese Texts: AI-driven Approaches to Linked Data Conversion

For rare books and special collections with classical Chinese materials, information interpretation and extraction for resource description is challenging due to the need for specialized knowledge of Chinese history and literature as well as the time required. The authors faced this challenge while working on a Linked Data (LD) project focused on describing Chinese historical places. The complexity of classical Chinese texts makes them difficult for humans to interpret, and even more challenging to convert into structured LD formats (subject-predicate-object triples) in English for enhanced data description, discovery, and sharing. These challenges motivated the authors to explore the potential of generative AI in processing classical Chinese texts and facilitating this conversion. Various generative AI tools available on the market were tested and the one works best with selected classical Chinese materials was identified. This presentation will share the insights gained from using generative AI, the experiences in refining prompts for improved outcomes, and the challenges encountered throughout the process.