Creating Rothko-Inspired Generative Art With a Quantum Computer
The above image is a procedural art piece generated with data from a quantum computer. Or just call it “generative quantum art”, if you prefer. It is an exploration of line and form as basic building blocks of art making.
But how does this work connect to Rothko? The reason I love Rothko’s work is because of his intense focus and appreciation for the most basic elements of art. His work can feel uncomplicated, but still striking and well-composed. He worked with basic shapes, but was willing to let them be imperfect. To me, this is the perfect starting place for making art with a quantum computer. Quantum computing itself is in its infancy and researchers are still figuring out what it can do. As an artist, I too want to figure out what a quantum computer can do. I want to create something simple, but something that can be imperfect like the current state of quantum computing. Rothko certainly never made generative quantum art, but I find his work inspiring and a good conceptual foundation for exploring generative quantum art. Rothko is famous for his Color Fields. Instead of color, I’m using data from a quantum computer. I call them “Data Fields”.
From color fields to data fields
For me as an artist working with technology, having an abstract understanding of how data is represented is quite empowering because there is no fundamentally right answer as to what data should look like. To a computer, all data is the same — 0 and 1. Even the data returned by a qubit is either a 0or a 1 (although it can be in a superposition of 0 and 1 before it is measured). It is up to us to choose how we represent that data. In many cases there are good reasons for representing data in a particular way. It makes more sense to represent audio data as sound, and it makes more sense to represent pixel data as an image. But technically, you could represent audio data as an image and pixel data as sound (and artists do exactly this). In my case, I am represent data from a quantum computer as vertical and horizontal lines.
It was a fairly profound moment to me when I finally internalized that data can look any way that you want it to, because it opens up the doors for us creative types to work in highly technical fields like quantum computing where artists generally don’t operate. This also helped me understand procedural and generative art. An image isn’t anything more than data, so the data can be manipulated and rules can be written that dictate the composition. In this way, the artist behind generative art is writing the rules that dictate the composition instead of creating the composition itself.
Personally, I’m excited by the idea of using something so new and so different to see what creative applications it can be applied to. And after reading this, I hope you are too. Quantum computing isn’t as out-of-reach as most people assume it is, and there are abundant creative ways to use this new technology today. For this Rothko inspired piece, I wanted to take an incredibly simple quantum circuit and draw the data on a canvas with some very simple rules as an exploration of line and form.
For the quantum portion, I chose the Bell State.
The Bell State
In quantum computing, the Bell State is the simplest possible circuit that exemplifies all the things that make a quantum computer unique. You can think of it as the “Hello, World” of quantum computing, and it looks like this.
It is made up of 2 qubits. We apply a Hadamard (H) gate to set the first qubit in a superposition — it’s in a state that’s both zero and one simultaneously, and could be equal to either zero or one when we measure it. We then apply a controlled-not (CX) gate between the first and second qubits for entanglement. Now, the two qubits share an intimate connection; their values are correlated, such that you can’t describe one qubit’s value without the other. We then measure the two qubits with the z-measurement gates.
When measured, the first qubit has 50% chance of being a 0 and a 50% chance of being a 1 because of the superposition from the H gate. The second qubit will do whatever the first qubit does because it is entangled via the cx gate. So the results should always be 00 or 11.
To create the art piece, I ran the circuit 1000 times on a quantum computer. The result was an array consisting mostly of 00 and 11 as would be expected. However, because quantum computers are still noisy, there were occasionally times when the two qubits didn’t agree, resulting in some instances of 01 or 10. This noise opens up some interesting creative possibilities. That’s it for the quantum part. Next I took that data and drew with it.
Drawing with quantum data
The rules for this piece are fairly simple. The canvas is divided into two 50 x 20 sections. The data from the quantum computer is then plotted into either of the two sections as a vertical line or horizontal line. The data is plotted to the top canvas if the first bit is a 0, and it is added to the bottom canvas if the first bit is a 1. Then, the second bit is used to determine the direction of the line. If the second bit is a 0, the line is horizontal. If the second bit is a 1, the line is vertical. Because the majority of the data is 00 or 11, that means that most of the data on the top section is a horizontal line and most of the data in the bottom section is a vertical line. The handful of times where a vertical line is up top or a horizontal line is down below is because of the noise in the quantum system. Personally, those errors are my favorite part and I like how the eye is so drawn to them. As quantum computers get better, there will be less error, which is ultimately better for the field, but not as much fun for art making.
With all the rules in place, it was time to create some art. I wrote the quantum portion with Qiskit and the classical computing with Processing. The source code for both can be found here. Because everything was generative, I could easily control and tweak the aesthetics, like making the lines thicker, longer, rounded, etc. I could also run the script on a number of different quantum computers to choose the output I liked. I could also choose to run on the same machine many different times. The results will be different every time, but because it follows the same rules, the composition will remain intact. The final composition above was run on the quantum computer called ibmq_paris.
I can’t know what Rothko would have thought about something like this, but I hope the connection here is salient. Rothko was a pioneer of abstract art. He was famous for many things, but his influence to me was his simplicity of expression. Like his explorations this piece uses the simplest quantum circuit and plots the data as simple line segments. And this is just the beginning. There is so much more to explore and I hope other creatives will feel empowered to do so as well. Quantum computing is complicated, but not intractable. It should be something that is approachable to artists with a little bit of technical chops and a willingness to explore. To me, a quantum computer is a paintbrush. It might be the world’s most complex paintbrush, but that doesn’t mean it should be out of reach. I look forward to someday seeing a thriving community of artists working with quantum computers as a part of their creative process.
Russell Huffman is a designer with IBM Quantum based in Austin, Texas. The above article is personal and does not necessarily represent IBM’s positions, strategies or opinions.
