With the Coronavirus outbreak, China being the worst-hit country, is still trying to come to grips with the Covid-19 with the help of Chinese tech companies who are at the forefront of this effort. One of such efforts is recently made by the country’s biggest tech company, Alibaba, who claims to have developed an artificial intelligence-based system to detect Coronavirus with 96% accuracy.
According to a news report, this new AI system has been trained on images and data from 5,000 confirmed coronavirus cases. It, therefore, can detect the virus in CT scans of patients’ chests with extreme accuracy against viral pneumonia cases. Qiboshan Hospital in Zhengzhou has been among the first to introduce this AI algorithm to aid clinical diagnosis — one of the badly hit areas with the highest cases of Covid-19. According to Alibaba, the system will also be adopted in more than 100 hospitals in the regions of Hubei, Guangdong, and Anhui.
Sign up for your weekly dose of what's up in emerging technology.
Alibaba, since the outbreak, has continuously contributed towards combating the virus — Jack Ma, the founder of Alibaba, had donated $14 million, earlier this year, in order to fight the disease. The company also donated another $144 million towards providing medical supplies to the worst affected Hubei province of China.
This new AI system, by Alibaba, was developed at the Academy for Discovery, Adventure, Momentum and Outlook (DAMO Academy), which was established by Alibaba in 2017 with an initial commitment of a global research and development initiative.
However, Alibaba isn’t the only one in the country pitching in to fight the virus — healthcare organisation Ping An has developed a similar AI-based system to pinpoint the disease from computed tomography images. Even search giant, Baidu, has also open-sourced its RNA prediction algorithm LinearFold to help researchers reduce the time taken to predict the virus’s RNA secondary structure.
AI-Powered Algorithm To Fight Against The Coronavirus
Earlier, nucleic acid testing was considered as the standard reference for diagnosing the Coronavirus as evidence of pathogenicity. But, since the Chinese National Health Commission has widened the criteria for the diagnosis, CT scan became the top method to ensure patients receiving standard treatment as soon as possible.
Researchers at Damo Academy told the media that the new AI tool is designed to distinguish between infected Coronavirus patients and the ones suffering from ordinary viral pneumonia with up to 96% accuracy by only analysing at a patient’s CT scan — in essence, the two have similar symptoms. Usually a human doctor, by comparison, takes up to 5 to 15 minutes to analyse a CT scan of a suspected patient, which include more than 300 images. However, the new algorithm can complete the detection process within 20 seconds, which in turn will ease up the pressure on hospitals.
According to reports, NLP natural language processing along with CNN convolutional neural networks are used to train the CT image recognition networks, which helps the system to identify the difference between Covid-19 images and ordinary viral pneumonia images. Besides, the system can also directly calculate the areas that are infected, thereby quantifying the severity of the cases to improve the efficiency of clinical diagnosis significantly.
This new diagnosis system is not Alibaba’s first attempt to combat the Coronavirus. Researchers from Damo Academy have also developed an AI-powered public health service tool that provides information related to the SARS-CoV-2 Coronavirus. The system can answer most inquiries regarding the pandemic via an app that serves as the Zhejiang governmental service platform.
Although Alibaba is popularly known for its online shopping and payment platforms, the cloud division of the company has been making bold strides in genomics research since 2015. The company provides computing power and data analytics to the Beijing Genomics Institute (BGI), one of the world’s largest genomics organisations.
Newer technologies like AI and machine learning have proved to be pivotal in addressing the coronavirus outbreak, according to WHO. Even other companies are helping scientists to battle the epidemic. In fact, Chinese scientific journal database — Chongqing VIP Information — made their academic papers free as an epidemic prevention effort. Alongside, donations of funds and resources have also poured in from tech giants such as Tencent, Xiaomi, Meituan and many more in response to the outbreak.