Google voice to text convertor – something you cannot possibly miss out using while you are using Google for any work. An android smartphone user would be quite familiar with the image below.
All the speeches that are commanded to the Google use data access to be converted into the text. In the process, our verbal command reaches the Google database and is stored there. Clearly, with the infinite incorporation that Google holds, it is certainly an example for Big Data.
For the beginners, I want to put forward the basic meaning of the Big Data. It refers to collection of datasets which is vast that all traditional statistical analysis fails and requires capturing, curating, storing, searching, sharing, transferring, analyzing and visualization with the help of advance techniques.
But then, I wondered that how is Google benefitted from this Big Data. The simple explanation that follows is that with the increased use of voice search and some YouTube speech transcription tasks, the size of language model keeps increasing. The major effect of the same is that error in transcription is reduced. According to a research paper published out of Google, the error rate that initially was as high as 37%-50% has been reduced by 6% to 10% with the Google’s Large Scale Language Modeling in Automatic Speech Recognition.
According to research director of Google Peter Norvig, the data size is directly corresponding to the model quality. He took a stance that as your data increases, the model learns more and eventually, the model reaches the optimization level. Even if the model was not a pearl of stuff to begin with, Big Data analysis and adaptation leads to model accuracy. Initially, the voice to speech problem was considered a linguistic problem which meant that a software would be required that has understanding of a human when it comes to interpreting one’s language. But Norvig improvised and designed statistical techniques and analyzed the compiled text that has already been translated. With the consumers seeking smoother and smarter user experience, each and every bit of data analysis becomes vital.
As the users now say- Google’s model of voice to text convertor has officially beat the Apple’s model viz Siri. The fact that Google receives more data to compute and analyze can be termed as a deciding factor in this role. This can be deduced from the fact the android users hold major market share at present.
Moreover, Google services are available across both the platforms while Siri is confined to the iOS. The formidable market share and across the platforms access allows Google to have vast Big Data and results are visible.
Comments