Machine Learning in Medicine - How AI is Fighting Covid-19
Is machine learning in medicine the key to beating the Coronavirus?
Science fiction regularly highlights the possible perils of AI. People like Elon Musk even say that it is only a matter of time until artificial intelligence eventually usurps control of society. However, right now in 2020, AI and deep learning in medical imaging are on the front lines of the fight against Covid-19.
What is Machine Learning in Medicine?
In March, the White House Coronavirus task force announced that it would be partnering with IBM, Google, and Amazon, to expedite Covid-19 research.
By pooling machine learning resources, the White House and major tech companies hope to expedite the development of a Coronavirus cure or vaccine. The only question is, how does machine learning in medicine work?
AI and Medical Research Basics
In principle, how machine learning in medical research works is simple.
By using predictive modeling, medical researchers can look at how a disease will likely respond to different forms of treatment. Researchers can also predict how pathogens like Covid-19 will spread throughout population clusters. This is why many countries are currently using tactics like social distancing to stall Coronavirus transmission.
Medical Machine Learning and Drug Discovery
Predicting how diseases might respond to different forms of treatment is itself nothing short of revolutionary.
Up until the advent of predictive modeling, many effective treatments for disease were discovered by chance. For example, when Alexander Fleming discovered penicillin in 1928, he did so only after finding a blob of mold in a petri dish, that seemed to be inhibiting the growth of a bacteria known to cause sore throats.
Even today, medical researchers rely on serendipity to discover new drug treatments. However, as revolutionary as discoveries like penicillin are, this means that new drugs can take decades to develop.
On average, it takes 12-years to identify new drugs. The development of optimal delivery methods can take even longer. Then trials need to be conducted to assess the safety of new medications. Now, though, with machine learning, drug discovery can take as little as a few weeks.
- By using AI and machine learning, pharmaceutical startups like Insilico Medicine can identify and synthesize new drug candidates in as little as 46-days.
- In January, it was announced that human trials will soon start on a new OCD drug called DSP-1181, which has been designed exclusively by AI algorithms belonging to British start-up Exscientia.
- At MIT, AI algorithms are currently being used to develop drugs like Halicin, that treat otherwise antibiotic-resistant strains of bacteria.
How Medical Machine Learning is Fighting Covid-19
One drawback to using machine learning in medicine rests with the fact that predictive modeling requires a lot of processing power. However, a quiet revolution in decentralized and cloud computing is helping remedy this problem.
Amazon, IBM, and Google aren’t the only big tech firms pooling tech resources to beat Covid-19. In fact, to aid in the fight against the Coronavirus, you don’t need to be a big tech company at all.
At present, anyone with idle computer resources can use tools like Folding@home, to participate in distributed Covid-19 research. This sees networks of millions of computers worldwide work together to run predictive modeling algorithms that analyze petabytes of real-time Covid-19 data.
Could You Help Fight the Coronavirus?
Millions of worldwide gamers, cryptocurrency miners, and everyday PC users, are currently helping fight Covid-19 using tools like Folding@home.
Could you too be part of the Coronavirus resistance? To find out, visit Folding@home or resources like Kaggle, to see how you too can contribute.