With the advent of winter in the northern hemisphere and the increase in the number of confirmed diagnoses in Europe and the United States, it is likely that more people will start to worry whether they contracted COVID-19 or became “asymptomatic infections”. In this regard, MIT scientists have developed an AI machine learning system that can be used to determine whether the cough is caused by COVID-19.
This AI is divided into three neural networks, one to hear the cough itself, one to check if the respiratory system is weakened (if the cough is getting weaker) and the third is to judge whether there is emotional instability due to neurasthenia. Although it seems very general, it has a high degree of accuracy in actual measurement. Among the samples known to have COVID-19, 98.5% can be correctly identified as sick. Among them, asymptomatic infections are even more likely. 100% recognition rate.
Scientists will turn this technology into an application as a “pre-screening” tool. As long as they hear it when they cough, they can understand the risk of illness a little. Obviously, this technology still has its limitations and cannot replace antibody screening or really quick DNA sampling, but it can relieve the pressure on the medical inspection system (and coughers) a little.