MIT is training an algorithm to detect Covid-19 via coughing

Contxto – What if I told you you could help stop Covid-19 by coughing… a lot? Bear with me here, there’s a tech-related explanation to this question, and not a sign of cabin fever (hopefully).

Researchers from MIT and the University of Guadalajara in Mexico are training an algorithm to “learn” to identify the sound of coughing of coronavirus patients. However, as with machine learning, the team needs people to literally “submit” the sounds of their coughs. That’s right reader, they’re crowdsourcing coughs.

Users interested in submitting audios of themselves coughing can do so here. Note that the process is done anonymously. At the end of the project, the researchers hope to give back by offering a free app.

[wd_hustle id=”InArticleOptin” type=”embedded”/]

Cough to fight Covid-19?

Besides your coughing audio, the platform will request you provide some additional information such as your age, sex, and if you’ve been diagnosed with Covid-19.

Even if you’re not experiencing cough symptoms, researchers equally invite users to force themselves to cough. Hey if you’re healthy, the algorithm will learn that from you. Because that’s the beauty of machine learning. As more submissions are made, the better the program’s accuracy to literally detect “the sound of Covid.”

The auditive power of machine learning has previously been used to initially diagnose Alzheimer’s disease, according to the research team’s leader, Brian Subirana.

For this Covid-19 case, Subirana says they’d need around “10 million coughs” in order for the algorithm to develop a 90 percent accuracy rate.

Related article: How Covid-19 created an Artificial Intelligence explosion

Once all the research is done, the team wants to launch a free app for the general population to use. Through it, they hope to reduce the spread of Covid-19 as self-isolation measures are relaxed.

To save a life and cough away… but not near others please. And clean up when you’re done.