Wednesday, September 22, 2010

Let the Light Shine Through: an interview with Adam Rokhsar


This is the first of (what I hope will be) a series of conversations/interviews with artists I like. I am proud to begin with Adam Rokhsar, who has been making some of the most inspiring videos I know (and uploading them on the web) since March 2009.

I wrote about Adam Rokhsar’s art about a year ago; you can read my post (which I still agree with wholeheartedly) here. Since then, Adam has continued making videos that proved my initial suspicion that he was a true artist. This is the link to his Vimeo profile.

I did no edits to his answers without his approval and had his final confirmation before posting this. Hope you’ll enjoy reading Adam’s answers as much as I did...





How would you introduce yourself?

I think of myself as an artist-scientist. This is a recent realization, that I could possibly be a synthesis of both and not forced to approach my work from one perspective at the exclusion of the other.

I have been going back-and-forth about what to write, if anything, to explain my work. As an artist, I want to create a space for people to experience the piece without my text intruding into their thoughts. I try to never intentionally create work that "ought" to be understood in any particular way; if I do, I throw that stuff out. I can't learn anything about myself or the world when I make work that has a meaning I know consciously.

However, as a scientist, I want to share what I learn from my investigations into the invisible structures of the world. If a visual effect is created by exploring some hidden pattern buried within a digital signal of music, I want people to know that. I want them to know they are seeing some relationship that I found, not made. It doesn't so much change what the work means -- it changes what the work is.

I struggle with these two perspectives. I created a piece recently in which the video content was pulled from the wall off the killers' apartment in the Hitchcock film "Rope." I took the digital video of a painting on their wall while I watched the movie, and then cracked open the video file. I broke the compression algorithm several times over. The video starting to look very slippy, liquid-like, but I could still see the figure inside the painting. Then I wrote software that analyzed a music track I wrote months ago, fed the results to a very simple learning algorithm, and that algorithm in turn chopped up the video even more. I wondered, should I share all that when I share the piece? I decided not to. I am still unsure of how to reconcile these approaches, and I think that speaks to an uncertainty about what I am still.


Which video is that?

I called that video Spirit is a Bone, Heaven is a Truck. Half Hegel, half Steve Malkmus.


Do you edit your videos?

I try not to edit the videos. If I do, it's just to cut off a little at the beginning or end, or if something really ruins the rhythm. I don't like to edit videos or music for that matter. When I try I tend to get compulsive about it unless I enforce a strict "first thought best thought" policy -- I might do an edit wherever the mouse happens to land me on the track. I trust accidents of my hand more than conscious editing decisions. I do my best to honor my errors as hidden intentions.


How do you think your past as "a therapist and behavioral science specialist" affect your videos?

When I was getting my graduate degree in counseling psychology from the University of Pennsylvania, my professor John Fantuzzo told my class that the job of the therapist was to "make the toxic invisible visible." I find myself obsessed with the same task as an artist -- to make invisible things become visible.

The other ways that being a therapist informs my current work are very hard to trace. During the six years I worked in mental health services, I met many people, saw many things, and encountered a variety of situations many people don't get to experience. It was a privilege, to be let into strangers' lives. I have counseled homeless families, runaway kids, juvenile sex offenders, addicts, pedophiles, victims of sexual abuse, developmentally disabled adults... It effected who I am profoundly, and must effect my work in ways I can't yet imagine.


There is a dark side to your videos, as if the discovery of the “toxic invisible” is bound to be horrifying at moments... Maybe that’s one of the answers? But it’s unlikely that your experience as a therapist is the only cause of this...

I don't think about my work in terms dark or light. It makes me so happy to make it, that's all I experience -- it's a wonderful feeling. After I finish a piece, sometimes I think, well, that was dark. I don't do it intentionally. I don't like to work when I know the conclusion ahead of time.

Whether the darkness you see is related to the traumatic material I encountered as a therapist -- of which there was plenty, especially when working with juvenile sex offenders and child victims of sexual abuse-- I'm not sure. I am interested in trauma, whether it's a force that pokes holes in our lives or whether its the part of reality our lives haven't covered, like the Lacanian Real traumatizing us with science and its hidden structures. I hope to make work that returns those structures back to the realm of human thinking, where they can become pure potentiality instead of a brutally unmovable pattern.


I notice that you use the word invisible a lot: “Invisible structures of the world”, “toxic invisible”, “invisible things”... You’re not the first at all but I find it interesting for a visual artist. In Joyce’s Ulysses there is a short poem:

I am the boy
That can enjoy
Invisibility.

Anyway, the “toxic invisible” is a way of looking at human psyche, while the “invisible structures of the world” sounds more metaphysical... These two aren’t necessarily related. Or are they?

I think you are right, that the toxic invisible and the hidden structures of the world are related. I like reading Zizek. He says, there are unknown unknowns, and there are unknown knowns. Things we don’t know we don’t know, and things we don’t know we do know. That’s the toxic invisible and the hidden structuring of things right there.

I never heard that poem, but I like it. Thanks for sharing it. I used to read Joyce a lot when I was a kid. My dad is a big fan.


On your website you talk about "augmenting our humanity with technology". How do you think we can go about doing that?

Right now I am really interested in neural networks. I don't like calling them that. How can you say something works like the brain when the materials are so radically different? A neuron is nothing like the set of computer instructions that make up an artificial neural network. The morphology of the two are fundamentally unlike.

That said, I am fascinated by what machine learning algorithms can do, and the way they are transforming our world. We have seen only the very beginning of it -- computer vision, natural language processing... where these things lead us are going to change what it means to be human, and I want to start using them right now to create art. So far the work I make with machine intelligence that isn't very different from other kinds of art. I am still making videos, images, and music. But I think there is potential for machine learning to allow for a radical shift in what kind of art is possible, in what art means. It is hard to imagine right now, but I am trying to imagine it. I want to make that happen.

It's important to me that artists get their hands on these tools. They are hard to reach. They require knowing programming, and math, and not so easy math at that. Plus they're mostly created by scientists and engineers who are not necessarily thinking of artists when they instruct others how to build machine learning algorithms. I would very much like to see these tools stolen and re-purposed to make art. Some artists do it now, but I know from my work as an art educator that to many people it feels out of reach. This is a problem. The new kinds of art are going to be very important for all of us, if we are to make sense of what it means to be human in the face of a new kind of intelligence.

I worry that we look at the machines around us and try to be more like them. I worry that we end up meeting the machines where they are and not the other way around. I think of the film "The Big Lebowski", when The Dude says, "Jackie Treehorn treats objects like women, man." I love that line. It gets to the heart of the issue: how we treat objects and people. I am not talking about intelligent robots. Look at the iPhone. It contains assumptions and rules encoded in it about what it means to relate to something. We push it, it does what we want. Clifford Nass and Bryon Reeves showed that people treat computers and machines as social agents, even if they know better. Their book came out nine years ago. Now I watch little kids flying through iPhones and games and wonder, what are they learning from these objects they treat as friends? Are they learning how to participate in mutually fulfilling, respectful relationships? What kind of relationship are these machine-human relationships modeling for us?


I’d like to hear more about the relationship between the “machine learning algorithms”, the “new kind of intelligence” and your videos... Can you be more specific? Please feel free to be technical...

I remember when I first watched the self-organized map (SOM) I programmed, well, organize itself. I had given it pictures of faces of my family, going back four generations. It was important to me when deciding how to program the SOM that I could see its learning process. I guess I wanted to get to know it, and I did -- I watched as the it learned what we all look like. It found the features that linked our faces in those particular photographs, I watched it move from its fantasy of my mom's face to my sister's, to my great-grandfather's. Strange but true: some researchers call the data that neural networks generate "fantasies." The way it moved made me immediately realize the accuracy of the term "Artificial Intelligence."

Since then I have become curious about what other learning models look like. I want to see the different kinds of intelligences people have been making. Right now, I am very interested in deep belief networks, and am getting a lot out of Geoff Hinton's research on them. Building a generative model is infinitely more interesting to me than building a model that only classifies, which I encoutered a lot in the Music and Audio Research Lab at NYU. Back then I was drawn to the hidden Markov model for exactly that reason; though it classified, you could always generate a state sequence and from there sample the gaussians to get new input data. The deep belief nets are more beautiful to me. I am eager to try and implement then in Max, which is where I built the SOM.


Can you tell us more about the programs you’re using? I downloaded the demo versions of some of them and tried to use but got lost very soon. I would have loved to join the course you’re teaching in New York. How can videomakers learn more about these? Any suggestions? Books?

I’ve been building everything myself in Max. My plan is to make the software I use free for anyone to download. Videomakers, musicians, people with no experience or interest in the technical side of AI -- this is my target audience. I really believe that no matter how complex these tools are, they should always be available to everyone and explained in such a way that people can understand what they are. Often I think programmers lose sight of how foreign their work is to non-programmers. I put a lot of time into making the software accessible, which often comes to do lots of visualization tools to make it clear what the algorithm is doing. When I finish the SOM library, I plan on posting in on my website and in the Cycling 74 forums.

As for books or courses, it’s hard to find a non-technical introduction to machine learning and AI. I can tell you, learning this stuff was slow going for me. I was always playing catch-up: with the math, the programming skills, the theory... it took a while and lot of work. I got good at reading articles I didn’t understand at all, and not letting that discourage me. I’m not above looking up every single word in a journal paper that I don’t understand, even if it takes a month to get through it.

I’m trying to address the lack of educational resources on AI for artists at Harvestworks, which is a non-profit arts center where I work as technical director. Within the next two months, I plan on running a course called AI for Artists, which will be designed to get people’s hands on new machine intelligence tools and give them an understanding on how the tools work without requiring four years of college math.


You talk a lot about technology but you use home videos and other personal things from your childhood as your starting points... Are they "starting points"? How do they affect the form of the videos? Is that part of the process of being transparent? One of your videos is called "The story I am trying to tell you..." Is there a story you're trying to tell us?

Family is definitely a starting point, that’s a good way of putting it. I have a lot of video material of my parents and sisters. They are very important to me and I think a lot of my work -- and who I am -- really has its roots in what it was like to grow up in my family. Our particular dynamics. I’ve noticed whenever I make a new program, the first thing I usually to do try it out is give it images of my family.

As I get older, I can’t help but feel a strange mix of wonder and sadness when I see the video footage of my sisters and I as kids. Computers make it easy to think that we can store the past in fixed form and bring it with us, but they can do nothing to address the fact that we cannot really cross back in time to become those memories. We can only have them. The tension between being and having is something a former therapy supervisor of mine taught me about. He believes that art is an attempt to reach back against the losses that come with time, the way your identity goes fractured, and make yourself whole again.


Why is your blog titled "Make Yourself Transparent"?

When I created Make Yourself Transparent, my goal was to make a place for my friends and I to share all the work we weren't finished with, wasn't quite ready to let out, etc... this is part of the idea behind the title. I decided to not wait until I was proud of my work. I wanted to just get it out there, in the hopes that I would spend less time thinking about how it would be perceived and more time making it.


A "part of the idea"... What’s the rest of the idea? And by the way, I’m not a big fan of your videos before Ontology...

I definitely agree with you -- the work before a certain point is simply not interesting to me. I don't think it feels like art. I only really felt like I started to hit a stride very recently. Maybe I will always feel like I am finally getting started. I think Faulkner said, you have to murder your darlings. Well, it's hard. I get attached to my work. So I try to put it out there and turn away. That's how I murder them. The less tightly I hold on, the more I realize I'm not really responsible for the work anyway. Then it's not such an ego boost. It's more about a way of living. I want to be transparent to the mystery -- that's what Joseph Campbell says. Be transparent to the transcendent. You don't make yourself into a light by trying to shine. You try to be very clear, don't be grabby, don't hold on to things. The light isn't mine. I am doing my best to let it through.


2 comments:

Jacob W. said...

really great stuff Yoel, thanks for this...

Yoel Meranda said...

glad you like it. and hope you like his videos...