The more it learns, the better it gets at going through the web for the content humans have always wanted.
This computer program learns from pictures on the internet by contextual phrase association.
Computer programmers from the University of Washington (UW) and the Allen Institute for Artificial Intelligence (AIAI) have created a computer program which is fully automated and they claim that it can “learn everything about anything.” The program is abbreviated as LEVAN.
It uses pre-designed image recognition and sifts through millions and millions of images found on the internet that are even remotely associated with a word or a phrase. Then it breaks these selections into categories and associates them with new compound words.
The results of the search spare users from clicking page after page of almost similar looking pictures. And the best part about it is that no human curator is needed. “The new program needs no human supervision, and thus can automatically learn the visual knowledge for any concept,” said research scientist Santosh Divvala.
There is just one drawback to it and that is the fact that there is a lot to learn for LEVAN. It has the vocabulary of a toddler presently and takes more than 12 hours to learn broader terms like “angry.” So, the researchers have invited the public to put their one-word concepts to LEVAN.