When Microsoft and Google are fighting for a better AI chat robot,this is not the only use of machine learning and language model.In addition to the rumored plan to display more than 20 AI driven products in this year’s annual I/O activity,Google is moving towards the goal of building an AI language model that supports 1000 different languages.In the update released on Monday,Google shared more information about the Universal Voice Model(USM),which Google described as the”key first step”to achieve its goal.

Last November,the company announced its plan to create a language model that supports 1000 of the world’s most commonly used languages,and also disclosed its USM model.Google describes USM as”a series of the most advanced voice models”.It has 2 billion parameters and is trained in 12 million hours of voice and 28 billion sentences in more than 300 languages.YouTube has used USM to generate closed captions,and it also supports Automatic Speech Recognition(ASR).This can automatically detect and translate languages,including English,Amharic,Cebu,Assam,and so on.

Now,Google says that USM supports more than 100 languages and will serve as the”foundation”to build a broader system.Meta is developing a similar AI translation tool,which is still in its early stage.You can read more about USM and how it works in the research paper released by Google.

One goal of this technology may be to be able to detect and provide real-time translation in the glasses of augmented reality,just as Google demonstrated in the I/O activity last year.However,this technology seems to be a little far away.Google’s wrong expression of Arabic during the I/O conference proves how easy it is to make mistakes.