How AI is helping us in Covid-19 pandemic?
Corona virus seems to have shaken up life as we know and inevitably it’s going to continue to spread but how far and fast that’s going to happen and what we’re going to do about it are still in question. So, could artificial intelligence play a vital part in providing some answers and maybe even improving the outcome? AI had an important role in the initial understanding of this outbreak.
Canadian specialists from BlueDot deal with global epidemic intelligence and identified very early on that something was wrong through a combination of medical and airline data. They were one of the first groups in the world that identified this outbreak. This was back in the morning of December 31st. The machine learning algorithms that they’ve developed had picked up information in China of a pneumonia wasn’t known to be a corona virus of unknown cause centred around this market in Wuhan. When it was presented to their team, they recognized that there were parallels to the SARS outbreak 17 years earlier.
And this is where the human scientists were needed. AI can alert but can’t yet do the investigation to say what’s really occurring. Detecting an outbreak is really just the first step, there are multiple other things that need to happen. Ultimately they need to determine what kinds of risks does this pose, is it likely to spread, if it’s going to spread where might it go, what kind of consequences might occur from that spread and then ultimately they have to get that information into the hands of people who can then be empowered with that information to take the appropriate actions.
It’s not just about tracking though the dream, it’s that one day AI might be able to conjure up necessary vaccines on the spot or repurpose drugs to deal with new challenges. BenevolentAI in London is already leading the way on. This company uses AI to better understand the mechanism of disease, correlating data on illnesses drugs and outcomes and providing more information than any human being could ever come up with and in this case a potential lead. They made an algorithm processing documents in real time, reading the abstracts of scientific publications to do with coronavirus and extracting relationships between key biological concepts that we really care about to carry out drug discovery. There’s a huge amount of information being published and they are not just pulling our coronavirus information but they’re merging it with a lot of existing publications on underlying biology. These algorithms reason over biology at multiple layers from the nano world of proteins interacting on ourselves through this maze of biological processes, up to human defined definitions of diseases.
Now a lot of people were looking at virals, antivirals and how they might potentially treat the disease. People from BenevolentAI looked at it from a completely different way. They were questioning what are the other types of approved drugs that might inhibit the progression of that disease in the body. So they surfaced a number of drugs and then did some experimentation based on that and further research. Then they came up with this one drug that they think is best suited. It has both anti-inflammatory properties as well as the ability to stop what’s called endocytosis which is what enables the virus to enter the lungs which is the most potentially dangerous outcome of the corona virus.
John Brownstein, chief innovation officer at Boston Children’s Hospital and a professor at Harvard Medical School and his team have built in a tool called Health map that gathers information from the US Centres for Disease Control and Prevention and the World Health Organization to provide the real-time data of the positive cases of Covid across the world. The project’s data is being used by researchers and governments all over the world. While the first vaccine for Covid-19 still has a long way to go before it’s available for clinical trials artificial intelligence is slowly helping us closing the gap on fighting the corona virus.
And what about China and their effort in fighting corona virus using the power of artificial intelligence?
Tech giants like Baidu and Alibaba are already at the forefront of the war by providing their expertise to the fight against Covid-19. In the wake of world’s attempt to prepare for the corona outbreak Baidu AI team has released a tool LinearFold to reduce prediction time from 55 minutes to 27 seconds. This efficiency is crucial for understanding the virus and developing its vaccine. The Covid-19 encode belongs to a family of enveloped corona viruses that are single-stranded RNA viruses such as HIV, Ebola and Influenza which muted faster and make vaccine development more difficulty. Alibaba has recently claimed that its new AI system can detect coronavirus in CT scans of patients chest with 96% accuracy against viral pneumonia cases.
14 days. That’s a time Chinese scientists took to identify the four genetic sequences of the novel coronavirus since its outbreak. it was shared to the World Health Organization immediately and has now become important foundation for global researchers in the combat against Covid-19. Now let’s turn the clock back 17 years. During the SARS outbreak it took months to identify the genetic material of the virus. From months to two weeks thanks to artificial intelligence playing an important role. Scientists first obtained BAL fluid samples from patients, isolated the DNA and RNA the sequence of genetic material by comparing them. AI’s advantage is ability to recognize patterns amid a deluge of data. Sequencing corona virus genome is like looking for needle in a haystack. Searching for genotypes with specific characteristics in a huge genetic pool. That’s where a branch of modern AI, heuristic search algorithms comes in. Combined with micro-genomic sequencing in the virus library this AI technology can also be used in tracking virus mutation. Think of it as efficient search engine. This search algorithm can shorten months-long process to weeks or even just a few days to identify a virus. AI doesn’t just support scientists, it can also help patients by improving the efficiency of diagnosis. China launched a genomic service which can run multiple virus gene comparisons and produce a detailed report within 60 seconds. This algorithm has been put into use in China and shared to the World for free.
Hospitals in China are equipped with an automatic temperature warning system and with people in China gradually going back to work AI temperature checking is also being used in public places like subways and office buildings. It uses facial recognition to locate one’s forehead then detects body temperature remotely through infrared thermography in just one second to avoid cross infections.
The world’s first intelligent AR temperature measuring glasses were also put into use during Covid-19 outbreak. The device can scan the body temperature of people in real time and show it to the wearer by way of thermal imaging. Up to 100 people can be screened within two minutes.
Using big data, speech recognition and semantic understanding technology AI can call people at the risk of infections. This “volunteer” can make ten thousand phone calls a day without taking a break. In Dongyun community in east China’s Suzhou City, limited staff are not able to screen all 7,000 residents but with the help of AI it can get a report with categorized information for their investigation in just two hours.
Now the part which makes the most controversy – Vaccines
Where is a vaccine for the novel coronavirus? According to the World Health Organization it is still 12 to 18 months away. it’s not a short time, but it’s surprisingly fast as it used to take 8 to 20 years for new vaccine. So how they achieved this breakthrough speed? AI played the key role. Vaccine research and development has three points: massive data analysis, literature screening and knowledge mapping. That’s exactly the force of artificial intelligence. To better illustrate, let’s first look at how vaccines work. Vaccines contain dead or weakened pathogens which can make the immune system to produce antibodies and they will remember how to fight a virus. The novel coronavirus is RNA wrapped in a coat of proteins which are what our immune system mainly responds to. But the jump from gene to protein is still a big one. This is where AI steps in. Using big data and probability models it helped researchers predict and select possible protein structures. AI is much faster and more accurate compared with human prediction. After obtaining these structures they can design vaccines to make them produce an effective response without making people sick. As the pandemic continues around the globe there is a strong chance that Covid-19 can be regular. If that’s the case, a reliable vaccine may be the ultimate solution for the world. The good news is that three vaccines have been clinical trials in China. What’s more Chinese scientists are working for International Cooperation to unlock the cure for the pandemic.
AI will play a greater role than it has already because what’s being done is genetic sequences are being taken, that is the understanding of the organism itself and this is put into databases that then can determine where a virus came from, where it might be going in the future. Of course, AI hasn’t solved the Covid-19 crisis but it has hopefully helped to some extent and should be learning from it to be even more helpful in any future disease outbreaks.