They jointly developed a deep-learning model which can predict the chances of coronavirus patients becoming critically ill in five, 10 and 30 days, according to a statement published by AI Lab.
Some COVID-19 patients were in a stable conditions in early stages, but their symptoms deteriorated at a rapid speed. The newly developed model can detect patients that will potentially become severely ill and help doctors conduct early-stage triage, the statement read.
It also noted that the new model using deep learning is more accurate in its predictions compared to traditional models that measure the severity of pneumonia.
Nearly 15 million people around the world have been infected with the novel coronavirus as of press time, according to data released by the Johns Hopkins University.
The research results were published in the Nature Communications periodical on 15 July.
The article was originally published in the Global Times.