One blog post By announcing this update, Google is trying to make its Assistant a more user-friendly key so that it will also be able to pronounce unique names and do speech recognition. Google said that now users can also teach the virtual assistant to recognize and pronounce unique names. The company said that now the assistant will listen to the users without recording them and will also remember them. Currently this feature is only available for the English language but Google will soon add other languages to it.
Another important feature is to recognize the user’s speech and extract the correct context of the conversation. For this, Google has reconstructed the natural language understanding model of Assistant. So that it can understand the context as well as improve the reference resolution. The company said that this update BERT Uses machine learning technology powered by. Due to this, Google Assistant will now be able to work at 100 percent accuracy for the command of alarms and timers. Apart from smart speakers, Google also talked about bringing this feature to other devices soon.
Google has also implemented BERT to enhance the quality of interaction between the user and the Assistant. Now Assistant will also use the old conversation to reply and provide feedback. At the same time, it will also be able to understand what is being shown on the screen. It will now try to make the conversation more natural by answering followup questions.