You can’t wait until all this is over. You look forward to the day when you will be allowed to go see your family and elderly relatives without worrying about their health; the day when you will book a flight which will actually take place instead of being cancelled just a few days after; the day when you will get all dressed up with somewhere to go, and you will no longer have to cover your face with a mask. You anxiously wait for the day when you will wake up and this dark Orwellian reality will be a distant dream.
But the truth is, and I think deep down you do realise that, this is not going to happen overnight. It’s not going to be a simple thing that just happens. Things will not be back to ‘normal’ on January 1 2021 as a whole genre of memes suggests. It’s rather something we have to work to accomplish. There are going to be ‘waves’, ups and downs, calms and storms. Until things slowly start to take a different shape.
With hospitals and healthcare professionals getting more overwhelmed by the minute, businesses struggling to make ends meet and infections surging globally, the urgent need for effective measures is still there. This is where technology is, once again, instrumental in saving lives. The danger which countries and governments were not prepared for, is pushing technology and artificial intelligence forward in the fight against the novel virus.
How do you fight something so invisible, so small but fast and smart…something that targets people if your army only consists of people? The answer is machines. Some are prejudiced against the automation of medicine (and everything else, really). Robotisation is still feared and considered an obscure invention which will bring no good. But what if, for the time being, this is the only way out? What if this is our only chance to outsmart the virus?
It is true that the adoption of computer modelling technologies is evolving at a slow pace and the sector itself is yet to be explored in depth and further developed. Furthermore, healthcare professionals have not yet fully embraced the full potential of artificial intelligence.
It goes without saying that machine-learning has to be perfectly precise when a human life is at stake. Machine learning algorithms can learn from data. However, system errors could lead to wrong diagnosis and patient injuries. At the same time, collecting patient data for the analysis could be difficult and many patients might find data-collection to be a breach of their privacy.
And yet, AI is being implemented in researching the virus — it has been from the very start. Here I have compiled just some examples of its application in the fight against the novel SARS-CoV-2.
AI In Detecting The Virus
COVID-19 was detected even before WHO could warn travellers of the emerging risk. Big Data and AI-driven algorithms developed by companies like BlueDot and Metabiota managed to predict the movement and potential outbreaks of the virus.
Through analysing data from reports, flight tickets and animal disease outbreaks, BlueDot was among the first to identify the risk of COVID-19 and predicted that the world is facing an outbreak of a new type of virus. On December 31 2019 BlueDot
“Identified undiagnosed pneumonia in Wuhan”.
The Metabiota strat-up predicted the outbreak in countries like South Korea, Thailand, Japan, Taiwan, Singapore and Hong Kong days before any of them reported their first case. It did so by utilising AI, machine learning and Natural Language Processing (NLP), and by studying human behaviour and levels of fear.
Furthermore, the company has built an Epidemic tracker platform which allows you to monitor not only the spread and number of COVID-19 infections, but also of 7 other pathogens on a global scale. Calling this “impressive” wouldn’t cut it.
AI In Drug & Vaccine Discovery
To understand the SARS-CoV-2 virus and its components, and to track its genetic mutation, researchers are using AI in the form of machine-learning models and systems. These help scientists to determine which components of the virus will provoke an immune response in the human body.
These models work with data. When enough data is provided to the model, it looks for patterns, thus helping the search for a drug or vaccine. The advantages of the automated learning include better accuracy and reliability of the findings. An example of AI implementation into vaccine research is the virtual screening of repurposed drug candidates and new chemical entities.
Machine learning enables the creation of models that learn and generalize the patterns within the available data and can make inferences from previously unseen data. —Archadi et al. in Artificial Intelligence for COVID-19 Drug Discovery and Vaccine Development
The term drug repurposing refers to using existing drug in the theurapeutic treatment of new diseases. It is the more cost-effective means of drug development because the testing has already been conducted and is precisely where AI technologies can and are being applied with their ability to study existing data.
A very vivid example is the ongoing EXSCALATE4Cov (E4C) project — a consortium supported by the European Commission’s Horizon 2020 tender for projects to fight the coronavirus and improve patient care. Led by the Italian Dompé Farmaceutici and developed by 18 participating institutions, the project utilises Exscalate (EXaSCale smArt pLatform Against paThogEns) — by far the world’s most cost-efficient and powerful supercomputer system.
Exscalate has data (also called ”chemical library”) of 500 billion molecules and a processing capacity of more than 3 million molecules per second. The objective is, through understanding the virus’ proteins, compounds and structure and through matching this structure to the molecules in the chemical library, to potentially identify a drug for the virus.
AI In Studying The Virus — Japan’s High-precision cameras ahead of the Olympics
Let’s talk about Japan where AI is applied to determine what measures should be imposed on visitors at sporting events and how many spectators should be allowed to attend the 2021 Olympic Games. With the competition being scheduled to take place in July 2021, the country is thoughtfully preparing and health and safety are being priority number one.
Last Friday an experiment was conducted with 16,000 people watching the Yokohama DeNA BayStars play the Hanshin Tigers. High-precision cameras, carbon dioxide-monitoring devices and wind-speed measuring machines were used to examine the crowd’s movements, as well as the movement of saliva droplets across spectators.
This data, combined with the findings of the Fugaku supercomputer, will help the Japanese government decide what measures will be imposed during the most important event in the world of sports next year.
Fugaku is the world’s fastest supercomputer. It has conducted simulations on the movements of airborne droplets and the effectiveness of plastic face shields and has established that almost 100% of the droplets escape the plastic shield.
Another Fugaku simulation visualised the movement of aerosol particles when people with no masks are sitting around a table. The simulation suggests that humidity lowers the risk of infection and limits the movement of aerosol particles. Fugaku, built by the RIKEN Institute and the Japanese Fujitsu tech company, can make 415.5 quadrillion computations per second which ranks it as number one among the world’s 500 fastest supercomputers.
AI Against The Infodemic?
Artificial intelligence has been and is changing the game in different departments within medicine even though it is yet to be explored further. Soon it will be a year since AI first detected the threat and it would be interesting to reflect on us as society, and on what we’ve learnt for the past (almost) twelve dreadful, challenging, dystopian, horrifying months.
It’s interesting how people perceived the idea of the unknown and the lockdown. On the one hand, it was an opportunity to explore the beautiful sides of human creativity. Some found in all this an inspiration and a source for creative ideas for their Tik Tok videos, a whole genre of memes, blog posts, social media challenges. They managed to find the silver lining at the eye of the storm. They expressed their compassion, solidarity and support for their communities and other struggling countries through music.
On the other hand, there were travel restrictions and accusations, debated safety measures and politicised masks, empty supermarket shelves and people having a row over toilet paper. The mass fear and panic were turned into hatred, suspicion and violence. Deserted streets, airports with no passangers and lifeless ghost towns. Fake news, rumours and conspiracy theories. The virus was called fake and non-existent. Bleach was suggested as a potential cure for the virus and people actually drank it.
We were, and still are, exposed to a staggering amount of (mis)information. People choose to believe what they want to believe. The information explosion has always been there. How can we then fight the infodemic which eventually makes the pandemic even more dangerous and puts more people at risk? Can artificial intelligence help here, too?
To an extent, yes — as long as the infodemic takes place within digital platforms, misinformation can be tackled and sort of controlled. But what about the rest of the misinformation which spreads offline and is basically untraceable? How can AI help? Can it help?
The pandemic has been the topic of many conversations I’ve had in the past few months. I’ve heard interesting and sometimes shocking points of view, some people ‘believe in the virus’, some people don’t. And this is precisely where the problem lies. People don’t know what to believe anymore. It’s enough for them to skim through a dubious article or two to form their own idea and theory about the origin and ‘realness’ of the virus. The confusion and false information spread fast, even faster than the virus itself. How do we handle that?
I think this battle is just as important because nonsusceptibility to the infodemic is an important tool in the fight against the pandemic. Not giving in to false information from unknown or uncertified sources is a step we can all take in order to save lives, literally. This is called social responsibility, this is what being considerate of your friends, family and yourself looks like. It’s not about AI. Things boil down to your own thoughts.
Yes, AI keeps on excelling. It gets smarter by the day, it reads, analyses, filters, it basically thinks. I think we could help it if we use our ability to think as well? Think about it.