Voice recognition: Mozilla is crowd sourcing to make AI work for the people

Posted by  theteleg   in       2 months ago     693 Views     Comments Off on Voice recognition: Mozilla is crowd sourcing to make AI work for the people  

With the voice recognition In the current AI boom, it’s obvious who has it: tech giants like Google, Facebook, and Baidu. Now Project Common Voice, a new project from the Mozilla (the nonprofit creator of the Firefox browser) is experimenting with an alternative to data monopolies, by asking users to pool information in order to power open-sourced AI initiatives. With the company’s first project underway,  Mozilla is sking volunteers to donate vocal samples to build an open-source voice recognition system like the ones powering Siri and Alexa.

Project Common Voice recruits web surfers to spend a couple minutes reading sentences aloud and/or listening to other people’s recordings to check their accuracy. The voice recognition is a very minimal contribution for volunteers: Just read three sentences to help the system understand what everyday speech sounds like. No need to go to a soundproof room or get a high-quality microphone. “We want the audio quality to reflect the audio quality a speech-to-text engine will see in the wild,” reads the projects FAQ. “This teaches the speech-to-text engine to handle various situations—background talking, car noise, fan noise—without errors.

“Currently, the power to control voice recognition could end up in just a few hands, and we didn’t want to see that,” Sean White, vice president of emerging technology at Mozilla, tells The Verge. Big companies have the benefit of filtering data coming in, but for companies like Mozilla, other methods need to be taken. “The interesting question for us, is, can we do it so the people who are creating the data also benefit?”

 Personalisation is a key factor Mozilla points out that sets its project apart from currently existing services. The company explains that traditional AI systems work on a collective datasets that doesn’t necessarily identify individual, smaller groups among them. This can skew an AI-based voice recognition software, favouring the popular and the majority voices.

 Although the AI systems we interact with on a daily basis are built on proprietary data, there’s a whole world of researchers and institutions publishing useful, if rudimentary, open-source alternatives.

To take these projects to the next level, though, proponents of open-source data may have enlist higher powers to take on the tech giants. Chris Nicholson, CEO of deep learning company Skymind, says, “We may need third parties to step in — NGOs, governments, and coalitions of smaller private firms— and pool their data.” Nicholson suggests that sharing health care data can improve medical imaging technology, and driver data can make autonomous cars more natural and intuitive on the road. Sharing these types of datasets, he says, “has obvious public benefits.”

JOIN OUR NEWSLETTER
Join over 3.000 visitors who are receiving our newsletter with the latest updates of Global News written by the very best of our News reporters/Writers
We hate spam. Your email address will not be sold or shared with anyone else.

About  

A blog scientist by the mind and a passionate blogger by heart ❤️. Fountainhead of TheTelegraff (Award winning blog), speaker at various international forums, and has done several ghost writing articles for international magazines and blogs. Life motto: Live while you can! Teach & inspire while you could & Smile while you have the teeth. Send me an email: thetelegraffnews@gmail.com