Event 2: MFA Exhibition

Having grown up watching those trippy videos by Cyriak Harris, I have at thing for the trippy, fractal-like, evolving movements. I'm inclined to believe that's why, when I entered the exhibit, I was drawn most to the piece by Hye Min Cho. Or perhaps her background as an electrical engineering major from Berkeley before pursuing an MFA here. I'm not sure.

The display had looping videos of these strange shapes.

It also seemed like people were more concerned with what the pieces looked like, or the emotions they evoked. I on the other hand saw more value in how the piece was created. After all, each of these displays had something to do with one of the objects she listed in the description.

MFA final project booklet.

This was on the wall next to the displays. It's about the size of a business card.

Somehow the displays relate to these things (wealth represented by cash, according to Cho), which in turn symbolize the artist's Asian roots. With some thinking one can guess at which object corresponds to which (e.g. I felt the right-bottom display, with its lattice-like structure, was reminiscent of a crystal lattice often found in minerals).

So I asked her further about how these funky shapes resulted from these objects.

Turns out it was a two-step process. In the first, Cho fed VGGNet, a deep neural network for processing and categorizing images, with photographs of each object that she took. This "trains" the AI, in what we commonly and vaguely call deep learning. Once VGGNet was trained, Cho had it render what the machine "thinks" such an object looks like. She then fed that output into a rendering algorithm she wrote, resulting in the shapes in the exhibit.

So in essence, this is what the machine "thinks" these objects look like. Cho didn't want to reveal which object was which, emphasizing that that's beside the point. And it really is. What I found really important was the questions about the future role of machines in artistic expression. Here we have machines not just assisting an artist's creative process, but rather an artist modeling the human thought process, having the machine be the "creative" instead of the human.

It's like the theory of determinism, which posits that free will is an illusion and the past can completely predict the future. If AI reach a point where it can completely model the human creative process, what will then become of art?

I guess we'll have to find out in the future.