AI tools and services are being used or offered by companies around the world to help fight the coronavirus pandemic. In a best-case scenario, whereby the virus transmission is massively mitigated, researchers from Imperial College London predict “there would still be in the order of 250,000 deaths in GB, and 1.1–1.2 million in the US” resulting from the coronavirus. Imperial College London’s analysis landed in Washington over the weekend and it’s said to be the reason behind the US stepping up its response. British PM Johnson warned that further measures in the UK will likely be introduced in the coming days and a coronavirus bill for emergency powers is making its way to the House of Commons. Much like in wartime, technologies and social experiments that under normal circumstances would take years or decades to be tested and implemented will be rushed into use in days or weeks.
Some AI assistance
China’s Tianhe-1 supercomputer is offering doctors around the world free access to an AI diagnosis tool for identifying coronavirus patients based on a chest scan. The supercomputer can sift through hundreds of images generated by computed tomography (CT) and can give a diagnosis in about 10 seconds. Alibaba Cloud has launched a series of AI technologies including the International Medical Expert Communication Platform on Alibaba Group’s enterprise chat and collaboration app, DingTalk. The platform allows verified medical personnel around the world to share their experiences through online messaging and video conferencing.
Another solution from Alibaba estimates the trajectory of a coronavirus outbreak in a specific region using a machine learning algorithm based on public data gathered from 31 provinces in China. Within China, it has a 98 percent accuracy rate. For researchers and institutions working hard towards a vaccine, Alibaba has opened its AI-powered computational platform to accelerate data transfer and computation time in areas such as virtual drug screening. Several of the other leading cloud players in China – including Baidu and Tencent – have opened up specific parts of their solutions for free to qualifying medical personnel. In the US, Microsoft and Google have also done the same.
Last month, scientists from South Korea-based firm Deargen published a paper with the results from a deep learning-based model called MT-DTI which predicted that, of available FDA-approved antiviral medication, the HIV drug atazanavir is the most likely to bind and block a prominent protein on the outside of the virus which causes COVID-19. In early trials, coronavirus sufferers are reportedly improving significantly using HIV drugs.
Hong Kong-based Insilico Medicine also published a paper in February which, instead of seeking to repurpose available drugs, detailed the use of a drug discovery platform which generated tens of thousands of novel molecules with the potential to bind a specific SARS-CoV-2 protein and block the virus’s ability to replicate. A deep learning filtering system helped Insilico narrow down the list and the company has synthesised two of the seven molecules and plans to test them in the next two weeks with a pharmaceutical partner.
British AI startup Benevolent AI has also been active in seeking to identify approved drugs that might block the viral replication of COVID-19. The company’s AI system examined a large repository of medical information to identify six compounds that effectively block a cellular pathway that appears to allow the virus into cells to make more virus particles. Baricitinib, used for treating rheumatoid arthritis, looks to be the most effective against the virus.
For its part, the White House has urged AI experts to analyse a dataset of 29,000 scholarly articles about coronavirus and use them to develop text and data-mining techniques to help scientists answer the following key questions about COVID-19:
What is known about transmission, incubation, and environmental stability?
What do we know about COVID-19 risk factors?
What do we know about virus genetics, origin, and evolution?
What has been published about ethical and social science considerations?
What do we know about diagnostics and surveillance?
What do we know about non-pharmaceutical interventions?
What has been published about information sharing and inter-sectoral collaboration?
What do we know about vaccines and therapeutics?
The entire COVID-19 Open Research Dataset (CORD-19) has been made available on SemanticScholar and will be updated whenever new research is published. While the outlook around the world is currently grim, some of these AI-powered tools and developments offer a glimmer of hope we may be to reduce the virus’s spread, improve treatment for patients, and ultimately conquer the coronavirus sooner than otherwise would have been possible.