Speech Processing for Disease – Prof. Ami Moyal, President, Afeka Tel Aviv College of Engineering – Voice Tech Podcast ep.076

Voice Tech Podcast 076

Episode description

Imagine if voice technology could be used to diagnose diseases! This could be a reality if voice tech is used to identify non-speech sounds, such as coughs. This focus is of particular interest at the moment as the world’s governments rally resources to protect populations against COVID-19.

This is one area of focus for this week’s guest, Prof. Ami Moyal, President, Afeka Tel Aviv College of Engineering, Israel. Prof. Ami also talks about the future of voice technology and what we should be teaching children for them to be successful in the world.

Teaching is the main function of Afeka, but it does focus on research and applied research, in collaboration with industry. This stimulates a culture of creativity.

The Afeka Centre for Language Processing has been researching the use of speech processing and artificial intelligence algorithms for providing a quick and readily available pre-diagnostic assessment of COVID-19 infection, without the need for human intervention.

When it comes to a rapidly spreading virus such as COVID-19, with millions of potential carriers amongst the global population, it is essential to identify likely carriers of the virus at the early stage of infection in order to prioritize testing efforts and break the chain of transmission.

Among the earliest symptoms of COVID-19 are vocal cord edema and vocal cord infection. These affect vocal cord patterns. Afeka is modelling samples of speech, coughing and breathing, from both symptomatic and asymptomatic carriers, to compare with models taken with health subjects. Afeka is also modeling vocalization of subjects that tested negative for COVID-19, yet are exhibiting similar symptoms. This will allow the differentiation between someone who is a carrier, and someone who is not.

Prof. Ami says that people were initially hesitant about using voice commands. But this perception has changed, even for simple voice commands. In the not-so-distant-future, we will be able to communicate with any object or machine using our voice. Machines will be able to communicate with humans, using simulated voices.

Taking Afeka’s COVID-19 research a step further, Prof. Ami says he imagines one day that our voices will be analyzed continuously by our cell phones, which will notify us in real time when to go and see a doctor, because it has discovered a change in our voice that may result in an ailment.

Google, Amazon, and Facebook have defined speech recognition as a strategic goal. This will lead to major advances in the use of speech recognition. Eventually, we will be able to communicate freely with any device, whether it’s our mobile phone, refrigerator, robot, or our car.

For content producers, they must think about how they are producing content for the search algorithms of the future. Searches will be instigated through voice recognition.

Looking to the future, we need to equip children with both the skills and the knowledge to use speech recognition technology. The current generation must be able to analyze data, and be able to solve unpredictable problems.

Quotes from the show

[08:28] It is essential to identify likely carriers of the virus to prioritize testing.

[08:37] Prioritizing testing breaks the chain of COVID-19 transmission

[10:17] Among the earliest symptoms of COVID-19 are vocal cord edema and vocal cord infection.

[12:02] In the future, we will be able to communicate with any object or machine using our voice.

[31:58] It is our duty to educate a generation of thinkers with skills that will enable them to achieve personal excellence.

Links from the show

About the author

Carl Robinson
Carl Robinson

Carl is the host of the Voice Tech Podcast. Since launching in April 2018, Carl has conducted scores of in-depth interviews with voice industry experts, building one of the most well known media brands in the voice AI space.

He also publishes Voice Chops Tuesday, the number 1 voice technology newsletter, enjoyed by thousands of voice tech fans each week.

Carl is a startup founder, product manager and data scientist, and recently presented a model for voice emotion conversion at ICASSP 2019.

To contact Carl, use the links in this box to send an email, tweet or message.

Share this article

If you found value in this article, please consider helping others by sharing to your network. Just click one of the links below.

What do you think?

0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Related Posts

Voice Tech Podcast 076
Jason Croyle Speak2Web
Rsz Girl 1990347 1280 Terry Rogers

Get notified about new articles

Upcoming Events


Featured Products