The global crisis situates itself perfectly to witness the extent Machine Learning (ML) can go to help us fight against an unprecedented threat like COVID-19. The pandemic has almost crippled the world and the outbreak hasn’t just been detrimental economically, it has taken so many casualties already. In a dire situation like this, technology is going to be used extensively – it is a challenge to render solutions when millions of lives are at constant threat. Here the primary use of technology is to make a way to a panacea. This piece will be covering the way ML is helping us fight the good fight against COVID-19.
Brett Averso and Andrew Satz – two graduates of the Data Science department of the Columbia University launched EVQLV, a start-up creating algorithms to computationally screen, generate and optimize millions of therapeutic antibodies. With the help of this technology, the duo is working closely to locate treatments for individuals affected by the coronavirus. The ML algorithms can screen for therapeutic antibodies that too with a higher success ratio. The process of discovering antibodies might cost us years of research and tests but these algorithms can easily locate the antibodies which can help us fight against the virus rapidly. “Studies show it takes an average of five years and a half-billion dollars to discover and optimize antibodies in a lab. Our algorithms can significantly reduce that time and cost,” says Satz.
EVQLV is performing computational discovery of antibodies and optimizing the same before sending antibody gene sequences to their lab partners. The lab technicians then engineer and test these antibodies. Although this could take them months, the process is still a better alternative as opposed to us taking years. The antibodies with higher success ratios are being moved on to animal studies followed by human studies. The duo believes that their technology can help them find closure to their discovery with a plausible treatment against coronavirus by the last quarter of 2020. Satz said that their algorithm is reducing the chances of drug-discovery failure. He also added that “We fail in the computer as much as possible to reduce the possibility of downstream failure in the laboratory. And that shaves a significant amount of time from laborious and time-consuming work.”
The antibodies that are being designed are to prevent the virus from attaching itself to the human body. To give a more nuanced understanding, Averso said, “The right-shaped antibodies bind to proteins that sit on the surface of human cells and the coronavirus, similar to a lock and key. Such binding can prevent the proliferation of the virus in the human body, potentially limiting the effects of the disease,”
With the rapidity by which COVID-19 is permeating at a global scale, a larger partnership is being offered by the joint venture of IBM, Rensselaer Polytechnic Institute and other US-based labs. This partnership is offering researchers access to AI-powered supercomputers to help in COVID-19 research. Averso and Satz, the duo are hopeful that these Machine Learning algorithms can catalyze the speed of discovering coronavirus therapies – develop and optimize the same. The duo says, “We are hard at work accelerating the speed at which healing is discovered and delivered and could not ask for a more fulfilling mission.”
The spread of coronavirus can primarily be located in the dearth of information regarding its early-stage symptoms. The scarcity of which has disabled people in recognizing if they are affected at all. So, if they travel from one place to another being oblivious, they are going to infect others.
But the governments are collating information of citizens such as their medical records and travel history with the help of big data tools. The data collation has helped them cut down on redundancies, scaling the data and streamlining it for structural benefit. This data collation has only been possible with the assistance of big data tools. The data gathered is then analyzed and visualized by the government which is helping them in drawing a plausible medical scenario. Data Science is also helping these institutions in scoping a tentative course to help them fight against the virus. The estimation is also to make beneficial assumptions – to make use of the resources that their infrastructure permits along with the budgetary constrictions that come along. With such estimations, the government is making arrangements to accommodate medical facilities and monetary resources for its citizens.
Can AI predict the next global pandemic?
AI comes with its set of limitations, and when it comes to their reliance on past data, it can be a passive tool to deal with any novel threat. So, AI can’t exactly tell us what entails the COVID-19 course. However, the machine learning experts from the Centres for Disease Control and Prevention could model the present infections but without the intervention of ML. Here they have used a method which they are calling “wisdom of words” where they are relying on what the normal people have to say about the pandemic’s progression. The researchers are then collating the answers to structure a model. Although ML models can’t be treated as preventive measures against the next global pandemic, that passivity is quite subjective. The model will give us a headstart to help us prepare better.
With Novelty Detection, machine learning technologies can identify and learn data or signals that weren’t formerly present. So, a model is trained on a healthy person who can help it figure the early-stage symptoms of a disease. So, if the model comprehends the source of a signal and where it’s leading them, the warning signs can be detected at an early stage. And, this model can be effective in case of a pandemic.
Hope is computable
As a collective, the society needs to acknowledge the things that it can predict and to add to that, ML has no reason to not find its application in predicting – that hope is computable and can be used to identify patterns of threats.
Comments are closed.