With the advent of the intelligence era, it will become the future direction of medical development to aggregate clinical data, clinical guidelines, omics data, and comprehensive knowledge data of the medical ecosystem, build a comprehensive medical brain, and provide help to clinicians, researchers, and public administrators. As the medical field has a strong knowledge attribute, and generally has high requirements for both quality and quantity of knowledge, the process of medical knowledge construction cannot be separated from the cooperation between machines and experts. To find the balance between the two, and maximize the ability of deep learning, greatly reduce the amount of human work, in the meantime ensuring the quality of data, is the key to quickly and high-quality construction of medical knowledge base. Based on this, iFLYHealth researches on how to accurately represent, locate, select appropriate knowledge, and design appropriate construction routes and technologies to enrich the knowledge self-learning framework which constantly build and update medical knowledge.

In the foreseeable future, knowledge self-learning will play a crucial role in semantic search, knowledge question answering, and clinical decision-making in the medical field.