2019 WAIC|"Medical AI+Rigorous Medicine" Is Highly Valued

2019-09-01 18:00:41 VHS 49

       On August 31st, the world-famous 2019 World Artificial Intelligence Conference (WAIC) came to an end. In the medical AI sub-forum with the theme of “path-finding medical AI, solving the problem of landing”, we  brought together professionals from various fields of politics, medicine, business and learning to discuss how to promote science and technology represented by AI through integration and change. Let technology truly empowered medical care and dreams come true.


       Since the development of medical AI in China, it was facing a deeper multidimensional challenge of exploration and refinement.


       Keywords for "medical AI + rigorous medicine"

       The application of AI in medical health must be different from other industries. Because diseases have complex risks and are vital to life, they must be very rigorous in their development. Professionals in various fields have repeatedly mentioned the following keywords:


  • Rigorous medicine. Zhang Qunhua, deputy chairman and secretary-general of the Internet Medical Branch of the China Medical Association, said that Internet medical treatment was the first half, and AI was the second half because of the rigor of medical treatment.


  • Clinical Experience. Yu Siwei, director of the Intelligent Hospital Construction Office of the People's Hospital of Guizhou Province, said that clinical knowledge was the most important compared to the data of artificial intelligence; Chen Shengyu, vice president of Philips Greater China, also said that AI technology must be combined with clinical scenes to be practical. What's more,  the data and standardization of the AI clinical library were very important.


  • Not enough standardization is one of the challenges. Wang Peijun, deputy director of Tongji Hospital affiliated to Tongji University, said that there were not enough landing plans and landing products because the data was not standardized enough, resulting in insufficient samples was not enough and the summability of the products was not high.


  • Knowledge map. Yu Siwei also proposed two methods of AI reasoning. One was that deep learning is a method of induction, and the other method was deductive, which was  a “knowledge map.” Only by breaking through this difficulty was artificial intelligence, which would be the next one difficulties and hot spots.


  • Start with the basics. Medical AI had great development prospects, but there were also many problems. You Mao, director of the data center of the National Health and Wellness Commission's Center for Health Development Research, suggested starting from a low-risk scenario and using AI and informatization to improve some repetitive tasks.


       Basic engineering of “Medical AI+ Rigorous Medicine”: AI self-diagnosis system

         In foreign countries, there were medical AI products that can fully satisfy all of the above keywords.

       For example, the Mayo Clinic, the nation's best hospital with a "patient-centered" mission, has long been at the forefront of the world's AI self-diagnosis system for common disease symptoms. Based on the rigorous attitude of evidence-based medicine, Mayo has developed a knowledge map composed of thousands of diagnostic decision trees based on more than 150 years of clinical diagnostic experience data, thus providing an online disease symptom self-diagnosis system for American nationals. More than ten years since its launch, it has been proven to effectively help American nationals to clear their medical needs, improve their health awareness, improve medical efficiency, and reduce the abuse of medical resources.

       In 2013, Mayo strategically invested in VHS and exported the system and massive medical knowledge map for the first time in the world. In 2015, VHS developed the AI self-diagnosis system, Mighty Doctor, using AI technology. Nowadays, through the earliest use of Mayo database knowledge, continuous domestic localization population adaptation and model transition, the system can answer more than 2000 common diseases and adapt to more than 95% of physical discomfort. More than 4 years on the line, landing in multiple scenes, it has helped more than 30 million people to get effective consultation advice:

  • Internet Medical: Embed Alipay in less than a year on the line, providing self-diagnosis for the majority of C-end users with scientifically rigorous and easy-to-use disease symptoms.


  • Insurance company: Embed the APP and public number of several insurance companies to provide daily health control tools for policyholders to improve medical efficiency and reduce medical expenses.


  • Community hospitals: As a pre-diagnosis and registration suggestion tool for patients, and a decision-making tool for doctors, it was used in many community hospitals across the country and was sinking to more cities.

"Mighty doctor" AI self-diagnosis process display