The competition for being the most used virtual assistant in the world starts from the struggle to be compatible with the largest possible number of languages.
When Apple bought Siri 10 years ago from a company – which had the same name – many wondered what it would be like to have a voice on our devices that listens to us at all times and responds to everything we ask. Total, if we already have the phone to find what we want, what advantages will something like this bring us?
But over the years and the improvement of technology, it has become clear that the leg of virtual assistants has become a huge battlefield on which more and more plots of current business models begin to depend. Who doesn’t ask Google Home to give you the weather, or to play a song and even buy movie tickets to watch a movie?
That fierce competition between Google, Apple, Amazon, Microsoft and Samsung is leading virtual assistants to play a crucial role in the current technological landscape where their success depends not only on how well they work, or the options they offer, but on how many Millions of people across the planet are able to interact with them.
How many languages does each assistant understand?
There are currently five great virtual assistants and they not all share the same market share. It is clear that Google has an advantage because it is present in most of the hundreds of millions of Android devices, while Apple is lagging behind with its iPhone, iPad, Mac, etc. The fact is that in that ranking where virtual assistants claim to be polyglots, the prize is won by Mountain View, with support for more than 40 different languages. Secondly we have the boys of Cupertino and his Siri, who can communicate with us in 21 languages.
Far from the first two positions, Microsoft appears with its Cortana, which at the moment is only able to understand eight languages. Do you miss Amazon? Quiet because Alexa is the fourth intelligence in this Top 5 thanks to the seven languages that support your Echo devices, Fire TV Stick, Smart TVs, etc … the same, by the way, that the assistant of Samsung Bixby – which is the one that is costing the most to open a gap since it is only present in the Korean brand devices and many of the languages he understands have not arrived until recently .
How difficult is it to add new languages?
It must be said that the addition of new languages is a colossal effort on the part of a company and that the current voice recognition systems are not easy to program. This is so because currently this whole process is left in the hands of deep neural networks that are able to predict phonemes and all those variations of the same word whose meaning depends on the context in which it is used. That type of learning is only possible because traditional voice recognition systems were left behind, in which the probabilities of word combinations were adjusted by hand, leading to a very restrictive scheme where, or things were said as expected the system, or we could not expect any response. These neural networks work even at a spectrographic level to reduce errors in interpretation, which eliminates human supervision in almost all cases.