DeepDream, the first music video created with Google neural network

Human bodies with animal heads. Mountains turning into buildings. Eyes scattered everywhere. These are not hallucinations, but the astonishing and even unsetting effects of the Deepdream technology, the open source code released by Google in July. This Artificial Intelligence System that uses neural networks to generate images through hidden algorithms, thus creating new dreamlike and grotesque

Human bodies with animal heads. Mountains turning into buildings. Eyes scattered everywhere. These are not hallucinations, but the astonishing and even unsetting effects of the Deepdream technology, the open source code released by Google in July. This Artificial Intelligence System that uses neural networks to generate images through hidden algorithms, thus creating new dreamlike and grotesque images, has been used to realise a music video.

[vimeo url=”https://vimeo.com/133275555″]

Just ten days after that Google published the code, researcher Samim Winiger used and implemented the Mountain View tool to create a video generator with which he made a music video of Calista & The Crashroots’s song Deepdream featuring singer Calista Kazuko. Even though this is an experimental work that Winiger doesn’t dare to define a work of art because “it is an exploration of generative systems”, the video aroused the interest of many artists. A few months later, indeed, the British band Years & Years asked the “code wizard” and his assistant Roelof Pieters to use the same technique, called inceptionism, to realise a video to their song Desire.

“We have an opportunity here to shift the paradigm of the way we interact with machines in film”, Brian Harrison, the director of Desire, said in an interview for Bloomberg. He was astonished at Pieters’ version of the film Fear and loathing in Las Vegas reinterpreted with AI. “I found it amazing that computers were thinking in a psychedelic way at first glance” he added.

 

Within a few weeks, Deepdream has become an innovative tool adopted by programmers, artists, designers and ordinary people who indulged themselves in making creative and creepy images autonomously. A technology that was firstly created on the sly and almost just for fun but that now is influencing art, videogames and, obviously, the cinema and music industries.

Translated by

Siamo anche su WhatsApp. Segui il canale ufficiale LifeGate per restare aggiornata, aggiornato sulle ultime notizie e sulle nostre attività.

Licenza Creative Commons
Quest'opera è distribuita con Licenza Creative Commons Attribuzione - Non commerciale - Non opere derivate 4.0 Internazionale.

Related articles
The best songs of 2017

Le canzoni più belle del 2017 secondo LifeGate Radio. Con questa playlist lanciamo la nostra collaborazione con Spotify Italia che vi farà ascoltare la musica migliore, selezionata.

Guatemala, Maya women fight to defend indigenous textiles from the fashion industry

An organisation that unites over 1,000 mainly Maya women in Guatemala has expressed alarm that indigenous handicrafts, textiles called “huipiles” in particular, are under threat because underpriced industrial fabrics appropriating indigenous patterns have flooded the Guatemalan market, depriving many native women of their main source of income. Read more: Down to Xjabelle, the fashion collection by a young designer with Down

RWANDArt, the documentary about young creatives putting Rwanda on the map

Rwandan creatives share a common dream: they want to use their art to bring about positive change in society, as well as put their country on the map. They’re mostly young, ambitious, self-taught and have a strong desire and determination to challenge mentalities, move forward and grow. In her documentary RWANDArt: A new creative generation, which