Alibaba's AI identifies Covid-19 infections with an accuracy of 96%

A new diagnostic system, using AI, promises to be able to detect new cases of Corona virus with an accuracy rate of up to 96% through computerized tomography (CT).

This diagnostic algorithm was developed by Alibaba's Damo Academy. Researchers at the institute say they have trained an AI model with sample data from more than 5,000 confirmed cases, adding that the system can identify differences in CT scans. between patients who have been infected with a new strain of the virus and cases of pneumonia caused by conventional viruses with an accuracy rate of up to 96%. According to the developers, their algorithms are integrated with the latest treatment methods, along with newly published research.

Alibaba's AI system will probably be deployed in more than 100 hospitals in Hubei, Guangdong, and Anhui provinces. It will help reduce the pressure on the medical staff in hospitals, because it can complete the identification process in just 20 seconds. It usually takes doctors 5 to 15 minutes to analyze a CT scan of a suspected patient and make a clinical diagnosis - however, scans sometimes include more than 300 images, causing Analysis and diagnosis time is very long.

The National Health Commission of China on February 5 expanded its criteria for diagnosing new cases, requiring CT scans combined with previous nucleic acid testing, to ensure patients with clinical symptoms will be treated properly as soon as possible.

This new diagnostic system is not Alibaba's first attempt at using AI to fight the Corona virus. Researchers from Damo Academy have previously developed a public health service tool using AI to provide information related to the Corona SARS-CoV-2 virus, which was first deployed by the Zhejiang provincial government. Giang on January 27. This system can answer almost every pandemic-related query through an application.

References: Nikkei