Short Bytes: Mozilla has launched a new open source project named Common Voice. It’s a speech recognition system that relies on online volunteers to submit their voice samples and validate the submitted samples. Mozilla aims to collect 10,000 hours of samples for speech recognition training.
Last year, Google said that about 20% of the searches made on mobile are voice inputs and the growth of voice searches is much more as compared to text input. Over the past few years, voice recognition and its use has become an important factor for a technology giant. Siri, Cortana, Alexa, and Google Assistant are its prime examples.
But, what’s happening in the open source world? Are these tech giants going to open source their speech recognition technologies and help developers build innovative products? I think you already know the answer. In order to create an open source speech recognition system, Mozilla, the maker of popular Firefox browser, has unveiled Project Common Voice.
As a part of the Project Common Voice, Mozilla is asking the volunteers to help train this open source speech recognition system. It asks you to spend a little time validating the sentences read by others and donate your own reading. The more voice you’ll donate, the better Common Voice will get.
Mozilla has set a goal of 10,000 validated hours of audio and plans to release the open source database later in 2017.
Companies like Google are already training their speech recognition and other AI systems using your data. They are able to convince you to do so by providing their “free” services. In Mozilla’s case, the speech recognition engine is open source too, so that makes even more sense to take a part in its development.
You can go ahead and give it a try here.
