Share on social networks
A few weeks ago, a rather strange image appeared, supposedly created by a computer, which pretends to show the result of the creation of a drawing with artificial intelligencefrom millions of images stored in his memory.
What this is really about is a research project that Google is developing, about neuronal networks.
Google affirms in one of its blogs that they have been working on a system of neuronal networks for the recognition of images through computational learning techniques.
Logr que una computadora llegue a un level of recognition of images by means of the denomination of “artificial intelligence” It’s very difficult, for that, Google implemented the use of computational learning and neuronal networks so that computers can recognize images of animals.
But in addition, they adjusted algorithms so that they not only recognize animals, but in any image, they recognize the animals that there are “apprentice” from others
“One way to visualize what is happening is to return the neuronal network and ask it to modify an input image, as a form of obtaining a particular interpretation”, mentioned by Google.
So, a team of Google researchers worked with a machine to detect animals in images. After that, it performed the process in reverse, passing to the algorithm any input image, asking it to recognize an animal and print the face it detected.
In this way, the computer is forced to create images with animal shapes on other images, generating results like the following:
“The results are intriguing – including a relatively simple neuronal network that can be used to over-interpret an image, such as when children enjoy looking at clouds and interpret the shapes. This network was mainly trained for images of animals, so it naturally tends to interpret the forms as if they were animals. But as the data is stored in such an abstract form, the results are an interesting mixture of these learned characteristics». text that can be read on Google’s blog.
The first image that was leaked to the Internet is the one that appears below, with which many theories arose about the domination of the world by machines, as predicted by Bill Gates or Stephen Hawking. All these are pure lies and aberrations of people who did not know what to invent about said image due to the lack of information.
Click on the image to enlarge
What is true is that if it is possible to perfect the level of computational learning in Google’s computers, it is unimaginable the capacity that computers have at the moment of creating “things”, because the images seen in the research blog are already a lot en qué pensar about artificial intelligence.
Original source of the article: Google Research Blog
Share on social networks
