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. This can be attributed to the unparalleled voice algorithm that the search giant has developed through the years, which was highlighted in an announcement that the Google Speech Team posted at Google's blog Thursday.
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. This technology is currently being used in the "deep learning" approach to artificial intelligence, according to Android Authority. It allows Google to better capture and analyze audio inputs, especially in noisy environments. The technology has been "trained" over the years by Google researchers to consume incoming audio successfully regardless of the disruptions caused by ambient noise.
The deployment of the deep neural networks allows users to have faster, more accurate and efficient voice search experiences, The Verge reported. The neural network acoustic model currently powers voice searches as well as commands for both Android and iOS platforms and voice dictation in Android. Researchers at Google explains that it is 300 milliseconds faster than the previous model.
