Google launched last year your voice recorder, capable of real-time transcriptions on devices like Google Pixels. As in practically all Google apps, this is possible thanks to machine learning, something that allows apps to be constantly evolving.
Good proof of this is that now Google has incorporated a new function called Smart Scrolling, another feature based on Machine Learning. Thanks to this, the most important words in the transcription are automatically chosen, so that they then appear in the toolbar scroll vertical. In this way, we can scroll through these keywords or touch them to navigate faster.
Google will recognize the most important words in the transcript
Google has just announced Smart Scrolling, a machine-learning-based feature for its Recorder app. With this new feature, the app will be able to choose the keywords you consider key in the recording, to highlight them when we are doing scroll in the transcript.
The learning model, according to Google, is so light that can be run offline, just from the device. The operation of this feature is quite interesting, as they have detailed in their blog.
This model consists of two main tasks: one extracts the keywords from each section of the transcript, and another chooses which sections of the text are most informative. Through different models of text extraction they can process all the tasks in parallel, analyzing the context of each word within the text.
But how can Google tell if a word is relevant? Using a learning model, the number of times the word appears in the text, the specificity of the term, that is, if that word has a wide semantic field or if it is very specific and other data about said word. In this way, the model is able to know the relevance of the word in the text, in a very summarized and simplified way (all this is much more complex, since it is pure computer science). Google has tested this feature on the Pixel 5 and it hasn’t given clues yet on when it will hit all devices.
Via | Google AI Blog
was originally published in