5 ways to use AI/ML to help mitigate the coronavirus outbreak
Perhaps we could look to artificial intelligence (AI) to solve some big pain points
Another day, another coronavirus article. I'm finding it hard to recall a time when we were not living in this dystopian reality we are trying to come to terms with. To ensure the number of cases doesn’t increase and the curve flattens, everyone's practising social distancing and adopting hygienic practices (such as wearing masks and using hand sanitisers). Most businesses have employees working from home, so at this point, we are hoping that most of society is living indoors and keeping themselves protected.
It is imperative to the survival of millions that we ensure that we practise social distancing until a vaccine is released, and history dictates that it could be a year to 18 months away. But, how can we effectively mitigate the virus's spread during this time and make our lives less complicated?
Perhaps we could look to artificial intelligence (AI) to solve some big pain points. Here's how:
1. Vaccine identification using machine learning
We're in the age where we can engineer our way out of a health crisis. The process surrounding vaccine development is tedious, and machine learning (ML) can help drug developers to quickly analyse suitable raw materials (organic, synthetic, etc.). ML models can examine protein structures, determine its impact on the coronavirus, and eliminate futile alternatives, thus helping researchers zero-in on the right substances for the coronavirus vaccine.
South Korea based Deargen has developed a deep learning model that can identify compounds that can tackle the coronavirus. The model has identified FDA-approved atazanavir (the HIV medication) to be able to cripple a prominent protein on the outside of SARS-CoV-2 (the virus behind COVID-19). Google's AI research lab DeepMind has open-sourced protein structure predictions associated with the coronavirus. Pathologists and researchers can use this data to zero-in on compounds for vaccine development.
2. Chatbots for COVID-19 diagnosis
We've already seen chatbots become the savvy medium for customer support. It answers queries accurately and solves problems faster than humans, thus reducing the load on customer support staff. Today, hospitals are reeling under unprecedented levels of stress, and chatbots can go a long way in helping doctors reduce the footfall they are currently experiencing. People who think they might be at risk could use chatbots to get themselves diagnosed and determine their probability of having the virus.
Chatbots could request body temperature, a list of symptoms, travel history, geolocation, etc. to evaluate the likelihood of a person being infected. As a result, individuals who are unlikely to have the virus can simply stay at home and avoid hospitals, thus relieving some of the pressure on these institutions. Apollo hospitals, a prominent Indian hospital chain, has developed a chatbot that can perform a quick diagnosis.
The Centers for Disease Control and Prevention (CDC) and Microsoft have created a chatbot that can determine the probability of being COVID-19 positive. World governments are promoting similar chatbots to ensure health, safety, and operational efficiency.
3. Predicting virus spread using ML and location data
Banks are already using ML to explore better ways to determine the default probability of an entity as well as fraud detection. It's also popularly used to create personalised marketing campaigns. Where there is data, there is an opportunity to deploy an ML model.
South Korea, the Asian country most affected by the coronavirus outside China, introduced an app that monitors citizens' whereabouts, thus enforcing social distancing. During a mandatory two-week self-quarantine, the app helped the South Korean government track the virus’ spread by identifying moments of contact (instances where a person was within two meters of an infected carrier.). The app helped South Korea sharply reduce its number of positive cases, thus showing the world that COVID-19 can be mitigated effectively. On the other hand, the app has also raised privacy concerns, since cheating spouses were exposed.
Israel has also developed a similar tool, a survey reading AI that can help mitigate the pandemic using user-reported data. Using this, populations can be warned about risky locations, and users at a higher risk of contamination can prevent being infected.
4. Hygiene enforcement using computer vision and thermal sensors
At a time when the coronavirus outbreak is causing a significant health crisis and massive economic havoc, governments could consider enforcing social distancing and hygienic practices (such as wearing masks and avoiding contact) using computer vision.
Using millions of training data images of crowded areas, computer vision cameras can be trained to detect neighbourhoods whose people might be at risk, and prevent mass contamination. Also, such areas can be disinfected on time, if necessary.
Enforcers can further mitigate the spread by using thermal sensors to evaluate whether people traveling from such locations are carriers of the virus. A Chinese firm has already developed a computer vision inspired product that governments could use to mitigate this pandemic.
5. AI systems in hospitals
Doctors face the highest risk in the era of this pandemic. Treating multiple patients in a day increases their chances of contracting the virus, and the last thing we need is a group of doctors unfit to practice.
However, AI systems are being developed to perform more contact-based medical procedures (such as body temperature checks using thermal sensors), and can automatically disinfect hospital areas that have a lot of human traffic (using sensors to determine human presence). Such systems can protect doctors and patients from increased contamination risk. Also, as a result, hospitals can test more patients and reduce the spread.
While we do our part to protect each other, we're also using AI/ML innovation at our disposal to tackle the coronavirus outbreak. Modern problems require modern solutions, and AI is here for us to embrace, as we soldier through these difficult times.
But, it's important to note that we're fighting an enemy we cannot see or know much about, so it makes sense for us to do what works: social distancing. It is the most effective control we presently have that will break the transmission chain and reduce the load on hospitals. It will also buy time for governments and AI models to identify treatments and build the infrastructure to handle the coming avalanche of new cases.
(Edited by Evelyn Ratnakumar)
(Disclaimer: The views and opinions expressed in this article are those of the author and do not necessarily reflect the views of YourStory.)