YOUTUBE wants more subtitle

The giant of the videos online: YOUTUBE wants more subtitle

Google video store exceeds 1 billion pieces with written description

Goggle at first had the idea to rearrange all the information that is in the internet, to make it accessible for all, to this day it is kept in that idea.


Under this idea descriptions were added to the videos.


The first videos stored on the Internet were stored there by 2016.



3 years after these came to light YOUTUBE.



From there until today they have reached the figure of 1,000 million videos with written description, called "CAPTION", managing to unite subtitles as a description of what happens.



The most relevant of this large amount is that behind it are not only people, which used to be the traditional method, rather it is technology.



Have made software not only recognize human language, they have also managed to learn about what is happening on the screen, differentiating whether it is in a daily scene like a class day, a moving car, or a Moment in a cafeteria.







According to the experts of YOUTUBE, every day they see around 15'000.000 videos with subtitles. English is the most requested language.




It has been evaluated that in the last year they have obtained an estimated progress of 50% of hits in their automatic system, although almost the same success rate has been obtained in the other languages.




But finding out who is behind this impulse, who is Liat Kaver, an engineer from COSTA RICA, is after passing through the MIT of BOSTON, has set out to demolish myths about the compression in the audiovisual world: "wants a Future in which everything contains a decription ".




With all this the more than 300'000.000 deaf people or with problems of hearing, can enjoy the videos,



Kaver remembers that in his childhood he did not find films that contained subtitles in Spanish, his native language.




Kaver felt he was losing a lot. Many times he intuited other times he created his own version in his mind.



Felt the need to create a system that would automatically display high quality descriptions.




That has been his goal in the past 3 years, and he has achieved it.





Although the progress has been remarkable, they still do not feel satisfaction with what has been achieved.



GOOGLE fully trusts that AI (Artificial Intelligence) will significantly help to make its goal scalable.




On the other hand they are aware of their mistakes and have cataloged them in lists, with the most comic errors, celebrations and the most sounded errors.





That they do as a cure to the Ego, always maintaining humility, humility that serves as a motivation to continue improving day by day.




"In the English language they have already achieved a percentage error rate in the transcription inferior to the one that we commonly have humans when we dedicate to do it"

Meanwhile the head of product says, "Our goal is to improve and make it work equally in the 10 languages they are developing"


These languages are Dutch, French, German, Italian, Japanese, Korean, Portuguese, Rudo and Spanish.




Due to the magnitude of the challenge, there is no specific date but to ask the contribution of the participants: "There is a community of creators and consumers who can edit the errors."