So I looked at a lecture by Gene Kogan. He is an artist that looks at generative systems and software as fuel for creativity. What intrigued me is that he’s writing a book called Machine Learning for Artists. Seeing that I’m finding it difficult to figure out how to merge ML and art, his lecture seemed like the way to go.
This piece is called Deepdream prototypes , it uses Google’s inceptionism code and artificial neural networks. in his own words, “The code accepts images as inputs and iteratively evolves the pixel values towards some coherent resemblance to the image classes it knows, producing wild images of “pig-snails,” “camel-birds,” “dog-fish,” and the famous “puppy slugs,” among many other categories.” I really liked it’s piece because it requires a sort of technical mastery with ML techniques as well as an aesthetic sense to produce the kinds of images you want to create. In my opinion it also lies in the uncanny valley because the resulting pieces resemble human art so much that it’s very awkward seeing the computational modification.
My favorite quote: He compared style transfers with machine learning to “Like if you rewrote the book of Genesis, Edgar Allen Poe style”