Google is working on optimising voice-based search to work even in less connectivity or zero connectivity zones. The solution is being built keeping in mind emerging countries, especially India where Internet connectivity is quite patchy. In an interview, Anjali Joshi, vice president of product management for Search at Google said that California-headquartered firm's engineers working on voice along with machine intelligence team have optimised voice to a point where it will be local to the device. "I hope to see it on the device rather than requiring connectivity. When that happens, we will see voice working a lot better."
Based in Mountain View, Joshi who is on an ongoing trip to the country said that the since India is the second largest nation in terms of all mobile searches globally, the company is focused on improving the search experience on the small screen and voice is an important component of it.
Google has also recently enabled offline access for many of its services such as YouTube in India to increase their usage in areas where connectivity is a challenge.
Talking about the backend technology, Joshi said that through deployment of machine learning and artificial intelligence, the computations are smaller to the point where they can be accommodated on the device memory of the phone and within the computation power of the phone. "This is a technology which is not just applicable to voice alone. Many of these machine learning, artificial intelligence driven models are being shrunk to a point where they can be done on the device. Once that happens, we will start seeing huge progress," said the IIT Kanpur and Stanford University alumni.
http://telecom.economictimes.indiatimes.com/news/google-india-now-adds-its-voice-to-slow-net-searches/51723227
Based in Mountain View, Joshi who is on an ongoing trip to the country said that the since India is the second largest nation in terms of all mobile searches globally, the company is focused on improving the search experience on the small screen and voice is an important component of it.
Google has also recently enabled offline access for many of its services such as YouTube in India to increase their usage in areas where connectivity is a challenge.
Talking about the backend technology, Joshi said that through deployment of machine learning and artificial intelligence, the computations are smaller to the point where they can be accommodated on the device memory of the phone and within the computation power of the phone. "This is a technology which is not just applicable to voice alone. Many of these machine learning, artificial intelligence driven models are being shrunk to a point where they can be done on the device. Once that happens, we will start seeing huge progress," said the IIT Kanpur and Stanford University alumni.
http://telecom.economictimes.indiatimes.com/news/google-india-now-adds-its-voice-to-slow-net-searches/51723227