DeepDreaming

English

The Mona Lisa with DeepDream effect using VGG16 network trained on ImageNet

Etymology

From DeepDream +‎ -ing.

Noun

DeepDreaming (uncountable)

1. The use of a convolutional neural network to find and enhance patterns in images via algorithmic pareidolia, thus creating a dream-like appearance reminiscent of a psychedelic experience in the deliberately overprocessed images.
   • 2015 August 3, Emerson Rosenthal, “Google's Deep Dream for Dummies”, in Vice‎^[1], archived from the original on 5 August 2021:

     #DeepDreaming is already as easy as a "click to upload" online tool.

   • 2017, Antonio Gulli, Amita Kapoor, TensorFlow 1.x Deep Learning Cookbook: Over 90 unique recipes to solve artificial-intelligence driven problems with Python, Packt, →ISBN, page 153:

     Many websites allow you to play directly with DeepDreaming. […] After the initial results presented in 2015, many new additional papers and blog postings were presented about DeepDreaming: […]

   • 2020, Neil Leach, “Do Robots Dream of Digital Buildings?”, in Philip F. Yuan, Mike Xie, Neil Leach, Jiawei Yao, Xiang Wang, editors, Architectural Intelligence: Selected Papers from the 1st International Conference on Computational Design and Robotic Fabrication (CDRF 2019), Springer Nature Singapore Pte Ltd, published 2022, →ISBN, page 61:

     The term “dream,” however, has been used—albeit metaphorically—in connection with a technique, DeepDreaming, discovered by Alex Mordvintsev of Google Artists and Machine Intelligence [AMI], while analyzing the operations at work in the process of recognition using artificial neural network.

   • 2022, “Interview: Matias del Campo”, in Dimensions, volume 35, A. Alfred Taubman College of Architecture and Urban Planning, →ISSN, page 089, column 2:

     MdC [Matias del Campo]: I am pretty sure it is probably the first built project that used DeepDreaming and style transfers as a design method.

Further reading

• DeepDream on Wikipedia.Wikipedia