Researchers from the University of Gothenburg, Sweden, created an artificial intelligence (AI) algorithm that could be used to contain the spread of epidemics, control outbreaks and find effective testing methods.
The current method to contain epidemic-level disease spread is random testing. This does not easily or quickly identify infectious people and leads to uncontrolled spread and outbreaks with more infectious people.
The team's algorithm improves on test strategies used during epidemic outbreaks with limited information. Its use could ultimately reduce the need for country-wide shutdowns.
A machine-learning algorithm was used to simulate an epidemic outbreak and information on the first confirmed case was analyzed to estimate infections in the rest of the given population. Data on the infected individual’s network of contacts, who they have been in close contact with, where and for how long were also used to estimate spread. The algorithm successfully brought outbreaks under control. In real conditions, more information could be added to improve the algorithm’s effectiveness.
Machine learning-based testing automatically adapts to specific characteristics of a given disease. It also has the potential to easily predict if a specific age group should be tested or if a limited geographic group is in a risk zone.
The team notes that the study was done in a simulation and testing with real-world data is needed to improve the method before it can be used in real life. This means that the algorithm is too early in development to be used for the COVID-19 pandemic, but they hope it could be used to prevent future epidemics.
This study was published in IOP Machine Learning: Science and Technology.