| Biography | |
|---|---|
 Prof. Qin Zhang Institute of Nuclear and New Energy Technology and Department of Computer Science and Technology, Tsinghua University, China |
|
| Abstract: DUCG (Dynamic Uncertain Causality Graph) is a new AI model and intelligent system developed for general clinical diagnoses to assist primary doctors lacking experience to correctly diagnose diseases differentially. Its causal knowledge bases including uncertainties were constructed by clinical experts instead of learning from big data. The knowledge bases aimed at patient’s chief complaints. So far, 36 chief complaints have been covered, including stomachache, breathing difficulty, etc. Most general clinical diagnoses encountered in primary hospitals and clinics have been included. It is because DUCG is based on causal knowledge that DUCG does not have the problem of generalization that the deep learning models have. All knowledge bases were verified by the third-party Grade IIIA hospitals with discharged cases recorded in the hospitals. The precisions were 95% or more, in which the precisions for rare diseases were 80% or more. The DUCG system has been applied in Jiaozhou City, Shandong, China and Zhongxian County, Chongqing, China. The diagnosed cases were more than 10,000 and near 10,000 respectively, in which only 6 cases were found incorrect because three diseases were not included in the corresponding knowledge bases (the two diseases have been included and the diagnoses become correct now). The cases are increasing every day. All knowledge bases and diagnostic results are interpretable with graph and text. The check recommendation function of DUCG can provide users with what next medical checks for an individual patient should be done step by step. The recommended checks were ranked according to the degrees considering dangers of possible diseases, comprehensive efficiency of each check to validate or invalidate possible diseases and costs of doing checks, while the ranked checks could be selected according to the local conditions. An online performance of DUCG of diagnosing disease will be shown. | |
| Biography: Qin Zhang is a member of Standing Committee of Chinses People’s Political Consultative Conference (CPPCC), emeritus member of China Association for Science and Technology (CAST), member of International Nuclear Energy Academy, fellow of China Association for Artificial Intelligence (CAAI) and director of the specialized committee for uncertainty in AI of CAAI, vice president of China Intellectual Property Society (CIPS) and director of academic committee of CIPS, professor of Institute of Nuclear and New Energy Technology and Department of Computer Science and Technology, Tsinghua University, former vice president of CAST, former deputy commissioner of State Intellectual Property Office (SIPO) of China, and former director of Science and Technology Commission, Chongqing, China. He published 20 papers as the first and corresponding author in journals ranking Q1 by JCR, and more than 100 other papers. He wrote a book: Basic Theory of Intellectual Property. He was grounded with 5 invention patents by SIPO and 1 invention patent by US Patent and Trademark Office (USPTO), while other 2 are under examination by USPTO. He winned the gold cup issued by International Federation of Inventor Associations (IFIA) during the international invention exhibition, Nuremburg, Germany, 2012. He presented a new AI model named as Dynamic Uncertain Causality Graph (DUCG) applied in (1) online fault monitoring, forecast, diagnosis, development prediction, decision support, and probabilistic safety assessment (PSA), for large and complex industrial systems such as nuclear power plants, chemical plants, satellites, etc., and (2) computer-added general clinical diagnosis, etc. So far, the precisions of DUCG diagnoses in more than 100 fault experiments of industrial systems have been 100%. The precisions of disease diagnoses for 20 chief complaints such as stomachache and breathing difficulty, for which the DUCG knowledge bases were constructed, were more than 95% verified by the third-party high level hospitals. These knowledge bases have been put into real use. The DUCG knowledge bases for other 17 chief complaints such as stethalgia have been constructed and waiting for the third-party verifications. The DUCG knowledge bases for another 7 chief complaints such as headache are still under construction. Hopefully, all the chief complaints will be covered and put into real use within one year, which means to cover almost all general clinical diagnoses. | |
