With the advancement of generative artificial intelligence,personalized learning tailored to individual needs has become an inevitable trend in modern education,and the personalized learning paths based on knowledge graph is the most common approach.Precise extraction of specialized terms is fundamental to the construction of knowledge graphs;however,there are problems with manual extraction:heavy workload,unintentional omission,and failure to update in a timely manner.The authors designed and annotated a dataset of medBaseDt,fine-tuned BERT,a large open-source pre-trained model,trained the termBERT model,and developed an in-telligent medical term extraction system.This system has been applied to textbooks of histology,embryology,and pathology,with an accuracy rate of 94.5±1.16％,exhibiting good effect.With this system,teachers can quickly obtain specialized terms from textbooks and create knowledge graphs.This technology also lays a solid foundation for subsequent construction of automated knowledge graph,intelligent test generation,and personalized learning.
关键词
基础医学/教学改革/人工智能/大语言模型/BERT/微调
Key words
basic medicine/teaching reform/artificial intelligence/Large Language Model/BERT(Bidirectional Encoder Representations from Transformers)/fine-tune
万方数据