Your phone tells you that you are infected with Coronavirus

 Researchers have used artificial intelligence to identify infection with the Coronavirus through coughing, so does this mean that in the future we may be able to diagnose ourselves through a short cough every morning in the smartphone?

The fact that smartphones know a lot about us and others is really nothing new. And a lot of people have long agreed that their devices and the app viewers that hide behind them know a lot about them.

 And for my smartphone to now recognize that I have COVID-19 - even though I didn't notice - that might sound like science fiction, but this is science, not fiction. Soon, a new application may appear on the market that will be able to diagnose infection with the Coronavirus through the sounds we make while coughing or talking.

While the success rate of the technique in those infected is good, the application must learn to avoid misdiagnoses in the uninfected. If this is successful, the application may complement the existing corona warning application, which identifies real contacts with infected people.

Three computer specialists from the Massachusetts Institute of Technology came up with the idea of ?? analyzing the sounds. Between April and May 2020, the three specialists recorded the votes of 5,320 of the test subjects. Besides coughing sounds there was language.

 The 4,256 voices of the test subjects were stored by the researchers in a computer that evaluated them thanks to an artificial neural network.

In this, the researchers bet on the vital properties of sounds that they had previously found in previous studies in patients with dementia. Then they experimented with what the machines had learned on the rest of the participants and the remaining 1,064 participants.

The outcomes of the trial were promising. "The model shows a sensitivity of 98.5 percent in test participants identified with COVID-19 in an official test," the researchers noted in their paper. And the characteristic thus reached in the group to 94.2 percent.

 This may mean that about one in twenty will get a false positive test result. In return, every person in five received a false positive warning.

 This means that the feature of the application must be improved in order for it to be usable, because if many people use it daily, false-positive results may cause panic and thus quickly lead to the exhaustion of the capabilities of the tests in laboratories.

In any case, the researchers concluded, "an inexpensive, effective and available mechanism at all times and in a large quantity for distribution to copy cases of Covid 19 and supplement the signs of disease containment."

Scientists suggest applying this test to students, students, workers, and employees every day to identify suspicious cough sounds. And in the event that a warning is issued, the concerned people can undergo an experiment in the laboratory to see if what the application has detected is correct.