With new improvements, the voice search tool is now faster and offers accurate results, even when used in noisy environments.
Advertisement
In the battle of personal assistants embedded in today’s mobile devices, it seems that Google, with its Now voice assistant, is pulling significantly ahead of its rivals: Apple’s Siri and Microsoft’s Cortana. These replaced the 30-year old standard in the industry: the Gaussian Mixture Model (GMM).
The result is that Google’s voice recognition software can now prove accurate even in loud environments, where every last part of a spoken word might not necessarily be clear.
The latest development in the way Google Now users interact with their devices through speech involves the so-called Connectionist Temporal Classification (CTC) and sequence discriminative training techniques, which improved on the Deep Neural Networks (DNNs) technology that Google adopted in 2012.
Our improved acoustic models rely on Recurrent Neural Networks (RNN).
While the sounds made by “u” and “j” can be a little hard to separate, the feedback loops helps the software to take all sounds and letters at once.
“In addition to requiring much lower computational resources, the new models are more accurate, robust to noise, and faster to respond to voice search queries – so give it a try, and happy (voice) searching”.
Advertisement
The new models are working in the Google app for iOS and Android, as well as dictation on Android, which works inside of some third-party apps, the team members wrote.
