When it comes to predicting diseases,
traditional statistical methods face two major problems: coverage is extremely
limited and statistical results are highly susceptible to subjective factors. That’s
why most approaches to predicting diseases are limited by the lack of data, and
even internet giants with massive data resources are struggling to overcome
this challenge. Liang Xu, Deputy Chief Engineer of the AI Department at Ping An
Group, which is also known as Ping An Technology, has accomplished a number of
innovative achievements in this area.
Xu has led the development of disease
prediction models in collaboration with the centers for disease control (CDC) in
two of China’s largest cities, Chongqing and Shenzhen. It is the first time that
massive city level data with over 20 million health and medical records were
used for disease prediction. The model can predict the epidemic status of
influenza, hand, foot, and mouth disease and the risk of chronic obstructive
pulmonary disease, with 90% accuracy. This model was highly praised by experts
from the National CDC, Shanghai Institute of Respiratory Diseases, and
government officials of Chongqing and Shenzhen.
Besides predicting diseases, Xu and his
team also applied Natural Language processing (NLP) techniques to predict life
expectancy by processing massive collections of health data and electronic
medical records. In addition, Ping An has successfully applied AI to assist
decision making in security and public services.
It took four years for Xu to build a team from scratch of over 80 AI scientists at Ping An, and he played a key role not
only in making general decisions and advancing the theory, but also in applying
the theory to practice and solving detailed technical problems. The success achieved
by Xu and his team has drawn a lot of industry attention.
In the future, Xu said he wants to bring the benefits of AI technology to more public
services to achieve the goal of improving more cities in China.