Google has developed an artificially intelligent neural network that’s transforming images into eye-popping pieces of psychedelic art.
Through a process Google is calling “Inceptionism,” the company’s research arm has developed image recognition software running artificial neural networks that mimic human brain function. Inceptionism then allows the software to analyze an image—its objects, layers and characteristics—and generates an alternative version of it akin to an impressionist rendering or a kaleidoscopic distortion.
Google engineers Alexander Mordvintsev, Christopher Olah and Mike Tyka introduced the Inceptionism in a blog post explaining how the various layers of the neural network could have great artistic potential beyond standard image recognition and processing.
“The techniques presented here help us understand and visualize how neural networks are able to carry out difficult classification tasks, improve network architecture, and check what the network has learned during training,” the engineers stated. “It also makes us wonder whether neural networks could become a tool for artists—a new way to remix visual concepts—or perhaps even shed a little light on the roots of the creative process in general.”
By “over-interpreting” an image, the engineers explained, the AI network iterates upon the original and creates impressions awash in color and “random visual noise.” The results, available in a Google Photos gallery, are what the engineers referred to as “neural net dreams.”