On the same day, The University of Cambridge posted a picture on Instagram of the campus' famous Corpus Clock modified by Deep Dream (image below, right) with the caption "We ran the #chronophage through Google's #deepdream algorithm. On July 3, 2015, Twitter user tweeted a modified version of a poster from the anime series Love Live! using the code (image below, left), with the caption "Oh lord what have I done #deepdream #lovelive receiving more than 170 retweets and 110 favorites. Between July 1st and July 6th, 2015, over 10,000 posts on Twitter have used the hashtag "#deepdream." Upon the release of the code, which was widely publicized, users began using the hashtag in droves to demonstrate their use of the tool. On July 1, 2015, the Google Research blog released the "DeepDream" code, created by Software Engineers Alexander Mordvintsev, Christopher Olah, and Mike Tyka in the Python language, as open-source via GitHub and encouraged users to share their image processing results using the hashtag "#deepdream." Some users created websites that use the DeepDream code to modify images, so as to create user interfaces for non-programmers or those not familiar with inputting images into Python code. On June 17, 2015, Google Research Blog's post "Inceptionism: Going Deeper into Neural Networks" described how the neural network AI analysis exploits certain features it recognizes and enhances based on images it has already seen, thereby creating figures in the image that Google referred to as neural network "dreams."
0 Comments
Leave a Reply. |