COVID-19, Artificial Intelligence
What AI can tell you about COVID-19 that you did not know before.
Separating the signals from the noise.
The COVID-19 situation is still dominating our lives. The “new normal” is a way to contain the spread of the virus, but it is not to last forever.
The scientific community is very active in studying the virus (SARS-CoV-2) and developing a vaccine or an effective treatment against COVID-19. Hence, many scientific articles are published about the virus, and it is not easy to keep track of them.
One way to efficiently explore those articles is to use the COVID-19 Open Research Dataset (CORD-19) that I learned about at an AI conference.
The site applies AI technologies such as Natural Language Processing (NLP), Text Mining, and others to handle the vast amount of articles that already exist, and that grows daily.
You get sophisticated search engines and filters to navigate your way through. You can personalize a feed to update you on new research, and you can learn about relations among the research done.
If you are interested in scientific articles, CORD-19 is a great place to start.
Another exciting site is Coronavirus Watch, powered by UNESCO, OECD, and other impactful institutions. The site provides access to worldwide media and advanced visualizations of coronavirus trends.
Medical research is also presented. You can search for relevant articles applying AI concepts, i.e., Natural Language Processing. The algorithms “understand” what the texts are about and offer you those that match your interests.
This concept is different when compared to ranking the articles based on an analysis of their titles. Those results would be biased (Titles aim to raise interest and are often not representative for the content as an analysis of the whole text based on NLP).
You may also like the visualization/simulation tools that provide you with a better understanding of the coronavirus and its spread in various countries.
There is also the possibility to embed the tools into your site. Depending on what you are working on, that could be an exciting option as well.
Racing for a vaccine, Artificial Intelligence plays a role, too. There is a lot of medical/pharmaceutical research ongoing worldwide, and researchers collaborate (and not merely compete).
The good news about it is that medical data can be pooled, and that pool can be analyzed as a conglomerate. The concepts that allow doing that draw from Big Data. Available data is, therefore, more condensed, and more knowledge can be generated out of it.
Implementing the above comes with some challenges. Records must be carefully mapped and aligned, and legal standards must be taken into account (as medical data is highly sensitive and hence highly protected worldwide).
Sometimes I get asked: With all that Big Data and AI available, why is the pandemic not manageable yet? The answer is: Big Data and AI make it possible to have a vaccine in place in a tenth of the usual time.
But of course, we are not there yet. And several assumptions about the availability of a vaccine are best-case scenarios.
However, it need not always be AI-enabled platforms to inform yourself about COVID-19. The emphasis, in any case, should be put on reliable sources. The two common ones are the dashboard run by John Hopkins University and the WHO.
