Communications of the ACM
Google Releases Dataset to Help AI Systems Spot Fake Audio Recordings
Two Google units have released a speech database containing "thousands" of phrases spoken by Google's text-to-speech models, in an attempt to identify bogus audio recordings.
Credit: Khari Johnson/VentureBeat
The Google News unit and the search giant's artificial intelligence (AI) research division, Google AI, have collaborated on and released a speech database containing "thousands" of phrases spoken by Google's text-to-speech models, in an attempt to identify bogus audio recordings.
The dataset is available to all participants of ASVspoof 2019, a contest for improving safeguards and countermeasures against falsified, AI-generated speech.
Said Google AI's Daisy Stanton, "Over the last few years, there's been an explosion of new research using neural networks to simulate a human voice. These models, including many developed at Google, can generate increasingly realistic, human-like speech. While the progress is exciting, we're keenly aware of the risks this technology can pose if used with the intent to cause harm. [That's why] we're taking action."
From VentureBeat
View Full Article
Abstracts Copyright © 2019 SmithBucklin, Washington, DC, USA
No entries found