----------empyre- soft-skinned space---------------------- Hi everyone, I wanted to start by thanking Renate and Tim again for inviting me back to Empyre. February was really great, I really appreciate the conversations—I had some excellent exchanges with Brian Holmes.
This topic is really interesting, and I am looking forward to the conference. The pairing of art, intuition, and technology is very timely, and I’m thinking about it now in relation to the dichotomy between astronomy and astrology Renate posed on Tuesday and in relation to co-creativity, which Jennifer mentioned in her post. As Renate mentioned, astronomy is the study and observation of astronomical objects, with the primary objectives being to record and mathematically predict their movements, and to potentially develop physics-based theories of causality. Astrology, on the other hand, observes celestial bodies in relation to our daily lives. Maybe I should first offer my perspective as an artist who works with technology like neural networks and stock trading algorithms. These technologies are sometimes are marketed as “advanced” or “cutting edge,” but I like think of them more as nascent, or even unfinished. In many cases, the popular conception of these technologies is primarily marketing hype when the realities are the limitations and “proper” uses of the technology are still open to debate. I first started college as a microbiology major, mostly because I really liked microscopes. I liked the idea of a device that could “reveal” something that has been there all along. I graduated with a BFA in photography—which is also a lens-based activity. I hadn’t reflected on this much until recently when I started to draw comparisons between working with neural networks to produce images and the work I used to do in a darkroom. In a darkroom, I would have a negative that I knew would produce some type of image. Depending on the filters I used, the length of time I projected it onto the photographic paper, and how I dodged and burned, the quality of the image produced could vary dramatically. In creating an image with a neural network feels similar in that I have a trained model, but the image it produces depends on the weights of the different parameters, how many epochs I have trained the model, and the images used in the training corpus. Like photography, these factors help to determine the quality of the image produced. When I was in the darkroom, I would experiment with less conventional techniques, like using multiple negatives simultaneously, or hanging the exposed photographs and painting the developer onto them and allowing it to drip down unevenly. The results were sometimes interesting, and could sometimes result in a better representation of the subject I was trying to photograph. But, these experiments also revealed something about the role of the photographic process itself. I am trying to do something similar now with neural networks and financial technology. How do you expose the role technology plays in the framing and creation of the meaning that it produces? Objective science usually takes the position that what is being observed is outside of the instrument being used to observe it—whereas an artist usually considers the instrument and what is being produced to be interconnected. This is what I feel is the strength behind the idea of co-creation that Jennifer mentioned. The role that technologies and social factors in the creation of meaning is quite often discounted. Neural networks are often described as “finding hidden connections” between different elements (people, things, economic factors, etc.). This makes a neural network similar to a microscope or telescope—which allow people to see objects that were either too small or too far away to be observed. However, I would argue that neural networks are more similar to cameras. They produce representations based on reality, but whoever is designing or setting the parameters of a the algorithm has a great deal of control over the version of reality that is produced. For an artist, this presents some very exciting prospects. But, knowing that this technology is primarily being used to help decide how to distribute resources, manage our retirement accounts, decide who can be released from prison early, and many other highly impactful decisions is very disconcerting because I don’t know if the scientists or technicians using this technology understand the role they and the technology play in the co-creation of meaning. I wanted to end by giving an example of some of the artwork Jennifer and I are making to try to reveal the role of neural networks. In a recent project, Going Viral, we used neural networks to generate videos of celebrities, politicians, and influencers who have spread misinformation about the coronavirus. The videos are public service announcements that are sharable over social media and correct the misinformation that was spread by the influencer. The shareable YouTube videos present a recognizable, but glitchy, reconstruction of the celebrities. The glitchy, digitally-produced aesthetic of the videos keeps them from being classified as “deepfakes” and removed by online platforms and helps viewers reflect on the constructed nature of celebrity and how neural network-based content recognition algorithms work. Here is a link to a short video that explains the project: https://vimeo.com/509818547 Looking forward to continuing the conversation, Derek -- Derek Curry, PhD. Assistant Professor Art + Design Office: 211 Lake Hall http://derekcurry.com/ On 3/16/21, 6:50 PM, "empyre-boun...@lists.artdesign.unsw.edu.au on behalf of Gradecki, Jennifer" <empyre-boun...@lists.artdesign.unsw.edu.au on behalf of j.grade...@northeastern.edu> wrote: ----------empyre- soft-skinned space---------------------- _______________________________________________ empyre forum empyre@lists.artdesign.unsw.edu.au http://empyre.library.cornell.edu
