Short answers, pulled from the story.
DeepDream is a computer vision program created by Google engineer Alexander Mordvintsev that uses a convolutional neural network to find and enhance patterns in images through algorithmic pareidolia. The result is a deliberately overprocessed image with a dream-like appearance often compared to a psychedelic experience.
DeepDream was created by Alexander Mordvintsev, an engineer at Google. The underlying network was built for the ImageNet Large-Scale Visual Recognition Challenge and originally carried the codename Inception, after the film.
Google released DeepDream in July 2015. The dreaming idea and its name became popular on the internet that same year.
DeepDream runs a trained neural network in reverse, adjusting an input image so that a chosen output neuron, such as one tuned to faces or certain animals, produces a higher confidence score. Repeating this process causes pareidolia, so the network generates psychedelic and surreal imagery even in pictures that did not originally contain the target pattern.
Researchers have noted that DeepDream's imagery resembles hallucinations produced by LSD and psilocybin, suggesting a functional resemblance between neural networks and particular layers of the visual cortex. A 2021 study in the journal Entropy found that DeepDream video triggered higher entropy and higher functional connectivity in participants' EEG signals, both recognized biomarkers of psychedelic experience.
DeepDream has been applied in the field of art history and was used for Foster the People's music video for the song 'Doing It for the Money.' A range of web services, mobile applications, and desktop software also emerged after Google open-sourced the code, letting users transform their own photographs.