Yidu Tech Decoded: “AI-powered super brain” continues to iterate, supporting the containment of COVID-19 in China’s cities

The prevention and control of COVID-19 is becoming a new normal, and the ability to rapidly identify virus transmission chains is the key to the timely control of the epidemic. In particular, for large- and medium-sized Chinese cities with over a million people, the need for balancing epidemic containment and promoting economic development presents practical challenges. Yidu Tech has been accumulating experience and technical capabilities since the early stage of the outbreak in Wuhan, and has successfully provided examples of how to effectively manage the epidemic for a number of cities in China.  

“City Immunization Platform”, an AI-based solution developed by Yidu Cloud under Yidu Tech, integrates functions such as monitoring and alert, simulation, trend analysis and prediction, control and dispatch command, and control policy implementation and tracking, forming an end-to-end AI-based epidemic combat system. Through data governance and data analysis, the system effectively assists in the prediction and analyzing of the epidemic and improves the decision-making efficiency of containment measures.

Dr. Yan Jun, Chief Technology Officer and Chief AI Scientist of Yidu Tech, said, “The monitoring and alert system must be capable of detecting the risk of an outbreak at its earliest instance, and provide multi-faceted sensors to enhance a city’s public health governance. In epidemic prevention and control, the analysis and prediction function acts as the core and brain of the operation, providing policy recommendations through data analysis; while strategy deployment and arrangement function acts as a baton, providing commands which ensure efficient policy decision and implementation.”

Data governance improved efficiency of containment measures

The "City Immunization Platform" supports effective data governance by allowing machines to automatically read epidemiological investigation reports and extract relevant fields through the use of natural language processing technology. The department concerned will then input data according to its authorization into the platform's decision-making engine, and through application of knowledge maps and machine learning algorithms, the virus transmission chains, relationship diagrams and timelines of different ID cases will automatically be generated and presented in a visualized format, effectively reducing the time and accuracy required for data processing. Through utilizing big data, the Yidu team have assisted relevant departments to analyze data and provide visualization to make accurate decisions. For example, in May this year, the system assisted the Guangzhou government to analyze data of 30,000 close contacts of confirmed cases, shortening the review time from the original two man-days to a mere 5 minutes.

Data and science supported epidemic analysis and prediction

Through the use of algorithms underpinned by "data-enabled intelligence with a focus of knowledge mapping and logical knowledge database”, the Platform evolved from the field of medicine to public health. By using big data to analyze the risk level and future trends of the epidemic, the Platform helped formulate policies for different levels and classifications of the epidemic. Our policy simulation model has the ability to simulate and predict the implementation effects of applying different combinations of policies. The simulation model would also make dynamic adjustments as different prevention and control policies are being implemented, allowing for more accurate predictions.

“AI-powered super brain” optimized through practical experience to support specialized research

Since the outbreak of COVID-19 a year ago, the Yidu team has supported containment of the epidemic in cities including Wuhan, Ningbo, Beijing, Qingdao, Shijiazhuang and Guangzhou. The Platform has been gradually improved from architectural design to function development through the accumulation of practical experience and algorithm training, becoming more intelligent and efficient.

Our solution is still rapidly iterating, with a focus not limited to COVID-19, but extendable to the field of public health management. For example, the Company has assisted Wuhan Tongji Hospital to develop a “COVID-19 intelligent registry”, which through the use of an algorithm-based engine helps hospitals improve data governance and identify accurate clinical research tools. The registry has helped more than 50 researchers or research teams including Wuhan Tongji Hospital to publish high quality articles and papers, with the highest impact factor of over 30.