On April 24, visiting Ph.D student Wu Mengcheng presented his study Knowledge Organization in Ancient Chinese Agricultural Texts in the NLP and Text Mining Seminar at KU Leuven.
This study implements natural language processing and knowledge engineering techniques. Drawing on a corpus of seven classical texts spanning over 1,700 years, the research applied entity recognition and domain-adapted language models to extract structured agricultural knowledge. This information was then visualized through knowledge graphs, highlighting patterns related to plant diversity, seasonal references, and disaster-related terminology. One notable insight was the recurring emphasis on the socio-economic role of silkworms and mulberry trees in classical agrarian discourse.