Construction of an artificial intelligence teaching assistant for obstetrics and gynecology residency training based on textbook knowledge structure
Xie Zhiwen
Liang Haixi
Zhou Lin
Lan Chunyan
Abstract：Objective To develop an artificial intelligence(AI)teaching assistant for standardized residency training in obstetrics and gynecology(referred to as"residency training"for short)based on the textbook knowledge structure and to verify its performance.Methods Taking Handbook of Obstetrics and Gynecology Residents,Clinical Guidance Manual of Obstetrics and Gynecology,and Clinical Skills and Clinical Thinking Series-Obstetrics and Gynecology Volume as core knowledge sources,the AI teaching assistant was built by structuring textbook content,establishing a self-developed question-and-answer database,and adopting a"retrieval-generation"dual-engine architecture(with ChatGLM-6 B as the basic engine).A total of 120 knowledge points were selected for performance verification,and the knowledge point accuracy,clinical logical rationality,clinical adaptability,and response efficiency were compared between this system and ChatGPT-3.5.Results The knowledge point accuracy of the obstetrics and gynecology AI teaching assistant was(91.2±0.33)%,which was significantly higher than that of ChatGPT-3.5[(64.5±0.35)%](P<0.001).Its clinical logical rationality was(87.1±0.42)%,which was better than that of ChatGPT-3.5[(63.8±0.50)%](P<0.001).Its clinical adaptability was(89.5±0.36)%,which was also superior to that of ChatGPT-3.5[(57.9±0.43)%](P<0.001).The response efficiency of this system was(1.7±0.4)seconds,which was significantly better than that of ChatGPT-3.5[(2.2±0.7)seconds](P=0.028).Conclusion The obstetrics and gynecology AI teaching assistant constructed based on the textbook knowledge structure has excellent performance.It can meet the needs of residents for fragmented learning and clinical ability improvement,and provide a feasible solution for the digitalization of obstetrics and gynecology residency training education.
Keywords：obstetrics and gynecologyresidency trainingartificial intelligenceeducational technology
Publication Date：2026-01-20
Online Publishing Date：2026-01-06(First online date of this platform, not the publication date of the document)
Pages：5( 21-25 )