摘要Chinese herbal traditions,with a history of over 2000 years,emphasise the harmony between spirit,body and nature.Integrating these principles with machine learning(ML)offers transformative potential for traditional Chinese medicine(TCM).By leveraging technologies and data-driven models,TCM can evolve while preserving its accumulated wisdom.Knowledge graphs combined with deep learning can enhance diagnosis,treatment planning and prognosis evaluation.This paper reviews current ML applications in TCM and strategies for integration with conventional practices.It categorises key challenges and pro-posed solutions,focusing on deep learning-based algorithms.ML has demonstrated success in automat-ing personalised herbal prescriptions,predicting diagnostic outcomes and identifying acupoints.However,major obstacles include data standardisation,ethical and legal frameworks,and fostering inter-disciplinary collaboration.The development of high-quality,ethical artificial intelligence requires regula-tory support and cooperation with TCM practitioners.This study supports the notion that a learning platform is essential for the education of TCM practitioners.ML and TCM may adopt this implementation approach,and the emergence of convex ML can substantially enhance testing algorithms in TCM,hence improving the effectiveness of future healthcare systems.