Artificial Intelligence can now draw images, based on your description


January 20, 2018 – FRN – 
GMbox – translated by Inessa Sinchougova 

Not so long ago, Google taught AI to draw so-called “doodles”, but until recently no program could draw something complex “on instruction”. However, a new technology from Microsoft has taught AI to better understand people’s requests and create images to your taste. 

Artificial intelligence programs have trouble distinguishing human speech. Because of this, they have many limitations, one of which was recently bypassed by Microsoft researchers: the company was able to mathematically describe and reproduce the human attention system.

According to researchers, it is thanks to the concentration of attention that people are able to create drawings. So, even such everyday things as “sharp beak” or “yellow wing” are inherently complex. Now that AI has learned to understand them, the Microsoft program can draw complex images based on your description. 

Furthermore, the system truly draws what it hears, rather than search for a corresponding image online. This allows them to create drawings of objects which do not exist. 

So far, the drawing skills of the program are still limited, but scientists are confident that this technology has a great future. 

Recently, AI have also beat humans at reading comprehension: 

The robots are coming for your job: A.I surpasses humans at reading comprehension

Subscribe to our newsletter
Sign up here to get the latest news, updates and special offers delivered directly to your inbox.
Notify of
Inline Feedbacks
View all comments
Would love your thoughts, please comment.x