Machines that learn human language in a way similar to how young children do

Machines that learn human language in a way similar to how young children do https://i2.wp.com/www.eresviral.com/wp-content/uploads/2018/11/Máquinas-que-aprenden-el-lenguaje-humano-de-modo-parecido-a-como-lo-hacen-los-niños-pequeños.jpg?fit=219%2C146&ssl=1

Machines that learn human language in a way similar to how young children do





Machines that learn human language in a way similar to how young children do


Very young children learn the language by observing their surroundings, listening to the people around them and connecting what they see with what they hear.



In computing, we use systems that are trained on the basis of phrases written down by humans, which describe the structure and meaning underlying that combination of words. These systems are increasingly important for web searches, to interrogate databases with natural language and for speech recognition systems such as Alexa and Siri. Soon, they could also be used for domestic robotics.



However, gathering enough of the annotated data can be difficult and lengthy for less common languages. In addition, humans do not always agree with the annotations, and in fact they may not accurately reflect how people speak naturally.







MIT researchers have developed a "semantic analyzer" that learns through observation to reproduce more closely the language acquisition process of a young child, which could greatly expand the computational capacity to understand human language. (Image: MIT News)



The team of Andrei Barbu, Boris Katz and Candace Ross, of the Laboratory of Computer Science and Artificial Intelligence (CSAIL), affiliated with the Massachusetts Institute of Technology (MIT) in the American city of Cambridge, has developed a system that learns through of observation, emulating the process of acquiring the language of a young child. This approach promises to significantly increase the learning capacity of the computer or robot that uses it. To learn the structure of the language, the new system observes subtitled videos, without other information, and associates the words with the registered objects and actions. Given a new phrase, the system can then use what has been learned about the structure of the language to accurately predict the meaning of that phrase, without the video.



This "weakly supervised" approach reproduces how young children can observe the world around them and learn language, without anyone providing a direct context.



The method could reduce the amount of work that specialists need to carry out to get a machine to learn human language.



Additional Information

        .

  !function(f,b,e,v,n,t,s)

  {if(f.fbq)return;n=f.fbq=function(){n.callMethod?

  n.callMethod.apply(n,arguments):n.queue.push(arguments)};

  if(!f._fbq)f._fbq=n;n.push=n;n.loaded=!0;n.version='2.0';

  n.queue=[];t=b.createElement(e);t.async=!0;

  t.src=v;s=b.getElementsByTagName(e)[0];

  s.parentNode.insertBefore(t,s)}(window, document,'script',

  'https://connect.facebook.net/en_US/fbevents.js');

  fbq('init', '369524843414444');

  fbq('track', 'PageView');

    .


SOURCE LINK ERESVIRAL.COM 
https://www.beviral.online