An early version of an app that can decide whether or not you’re affected by COVID-19 has been released. It provides the diagnosis by analyzing your voice. The creation is the brainchild of Carnegie Mellon University researchers working in collaboration with several other researchers from different institutions around the world through the internet.
Benjamin Striner, a Carnegie Mellon graduate student, involved with the project, said in an interview with Futurism:
I’ve seen a lot of competition for the cheapest, fastest diagnosis you can have. And there are some pretty good ones that are actually really cheap and pretty accurate, but nothing’s ever going to be as cheap and as easy as speaking into a phone.
Widespread shortages of testing kits pose a problem in some parts of the world, which makes a solution like this ideal for the current global coronavirus outbreak. And Striner believes that the algorithm used in the app could someday be a valuable tool in tracking the spread of COVID-19. For now, it is still highly experimental, and the team is again refining its accuracy by collecting more data.
Nevertheless, the app – called COVID Voice Detector – is already available. You can use it to analyze your voice for signs of infection, although the diagnosis is not to be taken seriously. The app comes with a severe disclaimer stressing that it’s “not a diagnostic system” and, therefore, shouldn’t be used as a replacement for a medical test approved by the FDA or CDC. Those organizations do not approve COVID Voice Detector, and it is still a work in progress.
Bhiksha Raj, a professor at Carnegie Mellon who worked on the project, said:
What we are attempting to do is to develop a voice-based solution, which, based on preliminary experiments and prior expertise, we believe is possible. The app’s results are preliminary and untested. The score the app currently shows is an indicator of how much the signatures in your voice match those of other COVID patients whose voices we have tested. This is not medical advice. The primary objective of our effort/website at this point of time is to collect large numbers of voice recordings that we could use to refine the algorithm into something we — and the medical community — are confident about.
If the app is to be put out as a public service, it, and our results, will have to be verified by medical professionals, and attested by an agency such as the CDC. Until that happens, it’s still very much an experimental and untrustworthy system. I urge people not to make healthcare decisions based on the scores we give you. You could be endangering yourself and those around you.
Striner has been working non-stop to prepare the app for release, and even he admits that it’s unlikely the program will ever be as accurate as a laboratory test. He said:
In terms of diagnostics, of course, it’s never going to be as accurate as taking a swab and putting it on some agar and waiting for it to grow. But in terms of very easily monitoring a ton of people daily, weekly, whatever, monitoring on a very large scale, it gives you a way to handle and track health outbreaks.
How It Works
- You download COVID Voice Detector onto a smartphone, tablet, or computer with a microphone.
- Once installed, the application is simple to use. It will inform you to cough several times, recite the alphabet, and then record a set of vowel sounds.
- Then you will be provided a score which represents how likely it is that you have COVID-19.
One of the other researchers working on the project, Rita Singh, has been creating algorithms that identify micro-signatures in the human voice. However, those micro-signatures have been linked to revealing physiological, psychological, and other medical data about the individual. But she has plenty of experience with making algorithms tailored to analyzing human sounds.
Singh, who is also a professor of computer science at Carnegie Mellon, said:
The cough of a COVID patient is very distinctive. It affects the lungs so badly that breathing patterns and several other vital parameters are affected, and those are likely to have very strong signatures in voice.
Striner’s team of ten Carnegie Mellon researchers have all been working on this project from home. The Carnegie Mellon campus is shut down due to the pandemic, and so are most institutions across the country. The situation has made it a bit more challenging to gather enough audio from confirmed COVID-19 patients to train the algorithm with.
That is why the team reached out to colleagues around the world to gather more data. Having to do that turned out to be a good thing because those researchers didn’t only send back audio from COVID-19 patients but from patients with other viruses as well so that they could train the algorithm to spot the difference. They even scoured news videos to find interviews with patients and included those in the dataset.
Striner said:
You have samples of people that are healthy, you have samples of people that might just have the flu. And you have all those different recordings of all the different types of coughs, like what are all the coughs that are out there? And then that allows you to kind of spot the differences.
Its accuracy cannot be tested currently because we don’t have the verified test instances we need. If it comes from a healthy person, we then have examples of what ‘healthy’ sounds like. If it comes from a person who has some known respiratory condition, we then know what that condition sounds like. The system will use all that data as counterexamples, and for disambiguating COVID signatures from those of other confusing conditions.
Since there isn’t a way yet to quantify the current version of the app’s accuracy, it isn’t a reliable tool to get medical advice from. However, the more people who use the app (healthy or otherwise), the more data the team will have to train the algorithm better.
