Chinese tech giant will provide data insights gleaned from its artificial intelligence-powered applications to help medical professionals combat current pandemic and carry out diagnosis more quickly.
Alibaba’s cloud business unit is offering its artificial intelligence-powered technologies to provide data insights on the current coronavirus pandemic and speed up diagnosis. The Chinese tech giant is hoping a suite of applications jointly developed by its cloud experts, researchers from its subsidiary DAMO Academy, and engineers from its communications platform DingTalk will support medical professionals worldwide in combating the virus. Alibaba said the AI-powered applications were put together from insights and lessons learned during the initial outbreak of the coronavirus. Pointing to its research arm DAMO Academy, it said three applications would be made available as a free trial to medical professionals and researchers, which could be used to aid in disease control and diagnosis.
The CT Image Analytics Solutions, for instance, is touted to improve testing accuracy and speed of detection in diagnosing COVID-19. Running deep-learning algorithms trained on data in China, the AI model can predict the probability of different pneumonia types, including the variety associated with COVID-19, according to Alibaba. Using lung segmentation methods, the system also is able to compute the proportion of lesions and affected volume ratio to the entire lungs. This process takes about 3 to 4 seconds to run and 15 to 16 seconds to transmit, which Alibaba says is almost 60 times faster than human detection. It added that more than 160 hospitals in China currently are using the CT Image Analytics Solutions.
In addition, its Epidemic Prediction Solution can map epidemic characteristics of COVID-19 in a particular region, generating estimates of size, peak time, and duration of the epidemic, and how quickly it may spread under three conditions: optimistic, neutral, and pessimistic. Powered by machine learning technology, the algorithm had been tested on data from 31 provinces across China, averaging at an accuracy rate of 98%. Alibaba is hoping this application can provide a reference to guide policymakers and medical researchers on prevention and control measures, medical resource allocation, and travel advisories.