PHOTO CREDITS: CHARLOTTE OBSERVER
According to the Massachusetts Institute of Technology, their engineers have developed an AI model that can detect the Coronavirus in a person’s cough with nearly 99% accuracy.
MIT scientists analyzed more than 200,000 forced cough samples, with their AI model accurately identifying 98.5 percent of people with COVID-19. This model can also detect the Coronavirus in people who are asymptomatic, just from listening to the way they cough. An asymptomatic patient has previously not always been detectable in any regard, though with this development of MIT researchers, asymptomatic and uninfected individuals are distinguished by analyzing recordings of coughs submitted by tens of thousands of volunteers online.
Voice recognition software works by breaking down the audio of speech into individual sounds, then analyzing each sound with a step of rules applied in order to spot the difference of coughing. This AI has many complex ANNs or ‘Artificial Neural Networks’ for prediction of a problem and for easy classification of those order of rules, or ‘Algorithms’. It can be simply said that if a type of tone is matched with the tone of a person who is not suffering from COVID-19, then that person is reported normal. If the tone of a person is matched with the tone of a person suffering from COVID-19, then that person is reported asymptomatic. That tone is broken down from a note of a person’s cough.
“The sounds of talking and coughing are both influenced by the vocal cords and surrounding organs,” says co-author Brian Subirana, director of MIT’s Auto-ID lab. “This means that when you talk, part of your talking is like coughing, and vice-versa.” The algorithm was able to find patterns in vocal cord strength, lung, and respiratory performance, and muscular degradation with uncanny accuracy. It correctly identifies all of the submissions from people who were asymptomatic but nonetheless tested positive for COVID-19.
“Since its main function is to distinguish asymptomatic coughs from ‘healthy’ coughs, this algorithm shouldn’t replace traditional testing,” Subirana added. This similar project was launched at Cambridge University, but it only reported an 80 percent success rate. This could be used internationally if just to know a suspected individual. Due to its extraordinary accuracy, it can be used regularly in many places. “Such a program could fight the spread of the virus if everyone uses it before going to a classroom, a factory, or a restaurant,” Subirana added.
ARTICLE: PATEL CHAITANYA
SCIENCE/HEALTH EDITOR: KYLE SMITH