CLOUDS Interactive Documentary

James George and Jonathan Minard co-directed an interactive documentary on the people in my scene. I was happy to be one of the interviewees and be one of the commissioned code artists. Here was the beta teaser on kickstarter:

I was responsible for implementing two of the open source visual systems: BallDroppings and Neurons. Being a part of this project was a very high honor for me and I only regret not being more available to involve myself during its production.

More info about CLOUDS on Creative Applications

and a good interview with them.

I'm posting this long overdue portfolio entry in celebration of CLOUD's screening at MoMa. In the tradition of these portfolio entries which share a small bit of the process, and indulging in the open-ness of the project, I'll describe my experience contributing.

We met up at a medium sized hackerspace in the Tenderloin district of San Francisco which was associated with the oooShiny tribe.  James, John, and a few others had already set up shop there when I arrived from Silicon Valley. A thin layer of ephemeral user network cables blanketed over the tables and floors, crawling at times like vines up the walls. USB, power, video, hot swapping in realtime to the crowd's murmur. It was a lot of macbooks, tripods, and hoodies. Exploratorium toys strewn about the space as they endearingly are year round. A dusty, cracked Processing textbook lay wedged between a copy of Curtis Roads' Microsound and a half-consumed makerbot filament case. I brought my cat Hexadecimal and let him poke around in all the boxes of miscellanea. People were walking around taking care of fifty different things. The space was teeming with life. The faint smell of smoke and curry wafting in through the victorian windows.

I joined in at the table with matching laptop and hoodie. James got me up to speed with getting all the git submodules. The platform was already organized and ready to go, all I had to do was subclass someone's handy superclass and fill it with my artwork. It wasn't just that ofxUI was already installed – Reza was in the house. The project was very deeply OF. I was also pleased to meet Patricio Gonzalez Vivo for the first time. Gmunk paid us a visit. I believe Grey Area Foundation was resident down the hall in the same space at the time. I think the whole project is a miracle.

Renoir paints RSAP protestors

artificial intelligence

This is a painting "by Renoir" of the protest group RSAP, or "Renoir Sucks at Painting". A new artificial intelligence deep learning algorithm that deals with artistic style was used to automatically piece different patches of Luncheon of the Boating Party onto a news media photograph of the picket line outside the Metropolitan Museum of Art. It took me over an hour using a GPU enabled EC2 instance to generate this image in 4 tiled parts. I'm especially grateful to the developers of neural-style for torch.

This generative piece is my personal artistic statement about RSAP, Renoir, and ai+art. I also tried using a screenshot of Habbo Hotel "Pool's Closed" as the style image, but it was less than recognizable, and this ai model seems overfit to paint strokes and everything else just looks paintstrokified.

Deepdream: Avoiding Kitsch

Yes yes, #deepdream. But as Memo Akten and others point out, this is going to kitsch as rapidly as Walter Keane and lolcats unless we can find a way to stop the massive firehose of repetitive #puppyslug that has been opened by a few websites letting us upload selfies. I don't think we should stop at puppyslug (and its involved intermediary layers), but training a separate neural network turns out to be more technically difficult for most artists. I believe applying machine learning in content synthesis is a wide open frontier in computational creativity, so let's please do what we can to save this emerging aesthetic from its puppyslug typecast. If we can get over the hurdle of training brains, and start to apply inceptionism to other media (vector based 2D visuals, video clips, music, to name a few) then the technique might diversify into a more dignified craft that would be way harder to contain within a single novelty hashtag.

Why does it all look the same?

Let's talk about this one brain everyone loves. It's a bvlc_googLeNet trained on ImageNet, provided on the caffe model zoo. That's the one that gives us puppyslug because it has seen so many dogs, birds, and pagodas. It's also the one that gives you the rest of the effects offered by dreamscopeapp because they're just poking the brain in other places besides the very end. Again, even the deluxe options package is going to get old fast. I refer to this caffemodel file as the puppyslug brain. Perhaps the reason for all the doggies has to do with the number of dog pictures in ImageNet. Shortly following is a diagram of the images coming from different parts of this neural network. You can imagine its thought process like a collection of finely tuned photoshop filters, strung together into a hierarchical pipeline. Naturally, the more complex stuff is at the end.

network visualization

What's the Point?

My goal in this post is to show you some deepdream images that were done with neural networks trained on other datasets – data besides the the entirety of ImageNet. I hope that these outcomes will convince you that there's more to it, and that the conversation is far from over. Some of the pre-trained neural nets were used un-altered from the Caffe Model Zoo, and others were ones I trained just for this exploration.


It's important to keep in mind that feeding the neural net next to nothing results in just as extravagant of output as feeding it the The Sistine Chapel. It is the job of the artist to select a meaningful guide image, whose relationship to the training set is of interesting cultural significance. Without that curated relationship, all you have is a good old computational acid trip.

The following image is a chromatic gradient guiding a deep-dream by a GoogLeNet trained on classical Western fine art history up to impressionism, using crawled images from Dr. Emil Krén's Web Gallery of Art. This version uses photometric distortion to prevent over-fitting. I think it results in more representational imagery. The image is 2000x2000 pixels, so download it and take a closer look in your viewer of choice.

deepdream by arthistory1 neural net

This one is the same data, but the training set did not contain the photometric distortions. The output still contains representational imagery.

deepdream by arthistory1 neural net

The below image is a neural network trained to do gender classification, deepdreaming about Bruce Jenner, on the cover of Playgirl Magazine in 1982. Whether or not Bruce has been properly gender-classified may be inconsequential to the outcome of the deepdream image.

High Resolution Generative Image

Notice that when gender_net is simply run on a picture of clouds, you still see the lost souls poking out of Freddy Krueger's belly.

High Resolution Generative Image

Gender_net deepdreaming Untitled A by Cindy Sherman (the one with the train conductor's hat).

High Resolution Generative Image

This was a more intermediary layer from deep-dreaming a neural network custom trained to classify various labeled faces in the wild (LFW).

High Resolution Generative Image

This was dreamt by the same neural net, but using a different gradient to guide it. The resulting image looks like Pepperland.

High Resolution Generative Image

This is the same face classifier (innocently trying to tell Taylor Swift apart from Floyd Mayweather) guided by a linear gradient. The result is this wall of grotesque faces.

High Resolution Generative Image

Just for good measure, here's hardcore pornography, deep-dreamt by that same facial recognition network, but with fewer fractal octaves specified by the artist.

High Resolution Generative Image

Technical Notes

Training neural networks turned out to be easier than I expected, thanks to public AMIs and nvidia digits. Expect your AWS bill to skyrocket. Particularly if you know about machine learning, it helps to actually read the GoogLeNet publication. In the section called Training Methodology, that article mentions photometric distortions by Andrew Howard. This is important not to overlook. When generating the distortions, I used ImageMagick and python. You can also generate the photometric distortions on the fly with this Caffe fork.

If you want to bake later inception layers without getting a sizing error, go into deploy.prototxt and delete all layers whose name begins with loss. In nvidia digits, the default learning rate policy is Step Down but bvlc_GoogLeNet used Polynomial Decay with a power of 0.5. I can't say that one is necessarily better than the other since I don't even know that properly training the neural net to classify successfully has anything to do with its effectiveness in synthesizing a deepdream image.

The highest resolution image I could train on the greatest ec2 instance turned out to be 18x18 inches at 300 dots per inch. Any more than that and I would need more than 60 gb of RAM. If anyone has access to such a machine, I would gladly collaborate. I also seek to understand why my own training sets did not result in such clarity of re-sythesis in the dreams. It's possible I simply did not train for long enough, or maybe the fine tweaking of the parameters is a matter more subtle. Please train me!