 |
| Artificial intelligence detects corona through coughing |
Researchers at the Massachusetts Institute of Technology have developed an artificial intelligence that can detect forced coughing in people infected with the Coronavirus, even when they are not showing symptoms.
The AI model is designed to detect the following diseases: pneumonia, asthma, and neuromuscular diseases, which change the way people cough in different ways.
Researchers have developed several neural networks that can be used to distinguish the subtle changes that signal the role of the new coronavirus.
One neural network recognizes sounds related to speech intensity, and the other hears emotions that reflect a deterioration of the nervous system (for example, increased frustration or a cold reaction).
Meanwhile, the third grid measures changes in respiratory and lung functions.
The researchers used an algorithm to check for muscle atrophy (that is, when the cough is weak) and get a more complete picture of a person's health.
Artificial intelligence is very accurate in the first tests.
After the research team used tens of thousands of cough samples to train their model, the technology identified 98.5% of coughs from confirmed coronavirus cases.
It also identifies people who appear to be 100% asymptomatic, and technology can provide a much-needed virus early warning system at these times.
The researchers said: We think this means that when you have the Coronavirus, the way you speak will change, even if you do not have symptoms.
This technology is not designed to diagnose people with symptoms as they may have other conditions that may lead to similar behaviors.
Scientists are working hard to develop an easy-to-use app that can be used as a primary virus detection tool. You may just have to cough on the side of your phone every day to see if you are infected.
The researchers suggest that if the tool continues to listen in the background, the technology may end the outbreak, even though it was a major event. Because it can lead to privacy issues.
The team is working with several hospitals to create a more diverse dataset, and is also working with a private company to provide an app that includes the tool for wider use if the company is approved by the FDA. he is. .