Am I the only one who found those images somewhat disturbing? I wonder if they're triggering something similar to http://www.reddit.com/r/trypophobia
Inceptionism: Going Deeper into Neural Networks
41–50 of 166 posts
Re: Inceptionism: Going Deeper into Neural Networks
#42The fractal nature of many of the "hallucinated" images is kind of fascinating. The parallels to psychedelic drug-induced hallucinations are striking.
I remember reading that drugs block certain parts of the optical nerve, which itself is a bit like fractal.
Also, personally, these pictures looked 'trippy' due to color palette that reminds hippy raves, rather than fractals.
Re: Inceptionism: Going Deeper into Neural Networks
#43The reason they look so 'fractal-like' (e.g. trippy!) is because they actually are fractals! In the same way a normal fractal is a recursive application of some drawing function, this is a recursive application of different generation or "recognition -> generation" drawing functions built on top of the CNN. So I believe that, given a random noise image, these networks don't generate the crazy trippy fractal patterns…
That's really cool.
The trippiness is further compounded by the rainbow-ish colour effect produced by the recursive function, which mimics the "shimmering" rainbow effect you commonly get around lights when tripping on LSD.
And also, when under the influence of various drugs you tend to see patterns, particularly faces, where there aren't any.
Re: Inceptionism: Going Deeper into Neural Networks
#44This is fascinating. And important. We need better ways to see what neural nets are doing. At least for visual processing, we now have some.
This might be usable on music. Train a net to recognize a type of music, then run it backwards to see what comes out.
Run on the neural nets that do face popout (face/non face, not face recognition), some generic face should emerge. Run on nets for text recognition, letter forms should emerge. Run on financial data vs results ... optimal business strategies?
But calling it "inceptionism" is silly. (Could be worse, as with "polyfill", though.)
Re: Inceptionism: Going Deeper into Neural Networks
#45The fractal nature of many of the "hallucinated" images is kind of fascinating. The parallels to psychedelic drug-induced hallucinations are striking.
If we want to progress AI, I seriously recommend these researchers try acid or magic mushrooms.
Re: Inceptionism: Going Deeper into Neural Networks
#46I would love to see what would come out of a network trained to recognize pornographic images using this technique. :)
That's a really worthwhile question, actually, considering the degree to which the human form and stylished portrayals of sexuality underpin so much of art and design.
Re: Inceptionism: Going Deeper into Neural Networks
#47The fractal nature of many of the "hallucinated" images is kind of fascinating. The parallels to psychedelic drug-induced hallucinations are striking.
Yes, this is basically artificial pareidolia. The interesting thing is that this technique would work on any kind of similarly structured neural net.
Re: Inceptionism: Going Deeper into Neural Networks
#48These paintings remind me of Louis Wain's work when he was mentally ill. Which makes me wonder, are these sophisticated neural nets mentally ill, and what would a course of therapy for them be like?
This sort of 'illness' should be supported, rather than treated.
Re: Inceptionism: Going Deeper into Neural Networks
#49Request for startup: Neural Network on demand artist.
E.g. SaaS that takes your images and use neural network transformations. Can you make a portrait of my that I look like king.
Re: Inceptionism: Going Deeper into Neural Networks
#50The one generated after looking at completely random noise on the bottom row, second from the right: http://googleresearch.blogspot.co.uk/2015/06/inceptionism-go... Reminds me very heavily of The Starry Night https://www.google.com/culturalinstitute/asset-viewer/the-st... Lovely imagery. I never had much luck with generative networks. I did some work putting RBMs on a GPU partly because I'd seen Hinton talk showing s…
https://lh3.googleusercontent.com/4jaIlDI1xXGOhxNejib833qA5y...
full resolution image