Artscape
A collaboration with the Intelligent Visual Interfaces Lab (IVI) at Rutgers University.
Conceptual artists like Sol Lewitt and John Baldessari presented instructions as art in the 70s.
Baldessari 1966-68 Acrylic on Canvas
Sol Lewitt for School of the Museum of Fine Arts, Boston, 1971
Proposal by Sol Lewitt submitted in 1970
Sol Lewitt, “wall drawing #260" (arcs and lines) originally produced in 1975
Sol Lewitt, “wall drawing #172" originally installed 1972, recreated in 2011 by Paula Cooper Gallery
Artscape conversely, uses a neural network to extrapolate "instructions" from something messy, complex and expressive. Abstract paintings and drawings are often inspired by words, phrases, fragments of thoughts and experiences. The neural network challenges this presumption as it has discovered a set of equations that approximate what's happening inside my brain.
So what is the Artscape?
The artscape is a collection of 1121 real and computer-generated artworks that are interspersed on a spatial grid. Of these, 400 are scanned images by Gautam Rangan over the course of years. They comprise a collection of sketches, illustrations, half-finished paintings, completed work, doodles.
The real art was used to train a neural network that learned my stylistic sensibilities.
Artwork below was hand-drawn by myself, then scanned and sent to the team at Rutgers University.
The digital artwork below was created by the Generative Adversarial Neural Network (GAN) at Rutgers.
My personal real scanned artwork below
Artwork created by the Generative Adversarial Neural Network (GAN) below
Sometimes the synthetic art fools even me into thinking it’s real. Can you distinguish the real from the synthetic?
Only the highlighted images below are real, painted by myself. The rest were generated by the GAN.
The Artscape Website
Artscape explorers start at a random location in an initially, barren artscape and begin their journey by minting art to reveal its neighbors . The art may be real or synthetic, unknown to the user at the time of minting. Each explorer crafts their own unique path while contributing to the collective unfolding and evolution of the Artscape.
An explorers journey culminates at the origin, manifested as a one-of-a-kind NFT which captures the path that led them there.
Example Journey Animation A
Example Journey Animation B
Example Journey Animation C
How Does the GAN work?
Below you can see two abstract artworks generated by the GAN and the 5 closest “neighbors” from the training dataset below. The training dataset is 800 of my hand-painted artworks scanned and sent to Rutgers University.
