Common Voice and DeepSpeech: Head of speech recognition leaves Mozilla

At Mozilla, work on speech recognition is apparently falling behind. Kelly Davis, director of the “Common Voice” project, is leaving the company. His last day of work was reportedly already this Friday. There is no question of a successor. Mozilla had just announced that it would have to lay off a quarter of its employees.

Mozilla wanted the personalities over first Golem had reported, did not confirm to heise online, but did not deny either. “Kelly Davis is collecting the greatest possible variety of voices for Mozilla’s Common Voice Project so that in the future everyone can talk to computers and be understood, regardless of their origin,” it says on another Mozilla website introducing employees. Davis worked in Berlin as the manager of the machine learning group.

It was only at the beginning of the year that Mozilla released voice control called Voice as an add-on in a beta version for Firefox. It should help you to find your way around the Internet faster and easier. Common Voice is a language database containing approximately 1400 hours of transcribed audio material. 42 speakers spoke texts in 18 different languages ​​for it. The data are freely accessible. The “Common Voice” project is being developed under the aegis of “Chief Innovation Officer” Katharina Borchert, who recently added another Award has accepted. DeepSpeech is an open source speech recognition.

After Mozilla had laid off 70 employees in January, a quarter of the workforce has now been added. According to CEO Mitchell Baker, the economic situation has gotten even worse in Corona times, so that the plans for 2020 are no longer feasible. All international branches are affected. The layoffs also bring with them a restructuring of the group, which is intended to serve competitiveness and sales. “The central insight is that our old model, where everything was free, has its consequences,” writes Baker.


To home page