7 ways to use machine learning against pandemic

Due to the health crisis, we're suffering, many researchers and scientists are collecting and sharing data, to learn from viruses and manage future pandemics. One of the technologies that are being used is machine learning.

 

How machine learning can help against pandemic: 

1. Identify who is most at risk.

Machine learning can help to predict the following risks: 

  • The risk for infection: What is the risk that a specific individual or group will be infected with COVID-19?
  • Risk severity: What is the risk that a specific individual or group will develop serious symptoms of COVID-19 or complications that require hospitalization or intensive care?
  • Outcome risk: What is the risk that a specific treatment will be ineffective for a particular individual or group, and how likely are they to die?

2.Diagnose patients

When a new pandemic comes, testing needs to be done on a large scale, which is complicated and expensive.

When it comes to using machine learning to help diagnose COVID-19, promising research areas include:

  • Using facial scanners to identify symptoms, whether the patient has a fever or not
  • Using wearable technology like smartwatches to look for significant patterns in a patient's resting heart rate
  • Use of chatbots with machine learning technology to evaluate patients for self-reported symptoms

3. Develop medications faster

When a new disease appears, it is essential to devise a vaccine, a reliable diagnostic method, and a drug for treatment, quickly. The methods that have been used conventionally involve a lot of trials and errors, which takes time. It can take years to find a viable vaccine.

Machine learning can speed up this process without reducing quality control. Researchers can find important molecules much faster. This has already been seen with other diseases like Ebola.

The researchers working on H7N9 found that ML-assisted virtual assessment and scoring led to substantial improvements in scoring precision. The use of the random forest algorithm provided the best results with H7N9.

In this situation, where we are battling a rapidly spreading virus, getting more accurate scores faster is crucial to speeding up drug development.

4. Find existing medications that can help

Companies spend a lot of time and money approving new medications. They should be as sure as possible that these medications won't have unexpected side effects.

This process protects us but also slows down finding the solution when we need a quick response.

An alternative is to reuse existing medications and use them to treat other diseases. The problem with this is that there are currently thousands of medications, and we don't have time to test them all, so how do we find the right one?

Machine learning can help us prioritize drug options much faster automatically:

  • Building knowledge graphs 
  • Prediction of interactions between drugs and viral proteins

5. Predict the spread of the disease

During a pandemic, when we are trying to develop strategies to work against it, we first need to answer questions like "How many people are infected?" and "Where are these people?" These questions are difficult to answer.

Generally, the government answers these questions, along with the health system. Since every day they count and publish the number of new patients diagnosed with the disease. The problem here is that there can be a big gap between getting the disease, developing early symptoms, and testing positive.

Fortunately, we live in a digital world. A person with difficult access to the health system can easily access social networks and have clues about their health.

It is possible that machine learning can classify people, it does not have to be at the individual level, but you can use all this data to estimate the spread of the pandemic in real-time and forecast the spread in the coming weeks.

 

6. Map where viruses come from

A zoonotic pandemic, like the one we are experiencing, is a pandemic caused by an infectious disease that originates from a different species and spreads to humans. These viruses can survive without being noticed in the natural world for a long time, waiting for the next mutation and the next chance to infect us.

Knowing the hosts that allow them to spread is vital to fighting a pandemic. Once they are found, we can develop strategies to control the spread of the disease and prevent further outbreaks from occurring.

Thanks to huge technological advances, the whole-genome sequencing, the process of determining the complete DNA sequence of an organism, has become cheap and fast. Research has shown that machine learning models can use genome sequencing data along with expert knowledge to identify the species that act as hosts of the virus.

7. Predict the next pandemic

Accurately predicting whether a strain of influenza will leap from one species to another can help doctors and medical professionals anticipate possible pandemics and prepare accordingly.

With machine learning, the researchers were able to identify potentially zoonotic influenza strains with high levels of precision. More work is needed to establish prediction models for direct transmission, but knowing which influenza strains can leap is an important first step in preparing for the next pandemic.

on May 7, 2020