The Art of Engineering and Engineering Art

Yes, that's a house. Is the art good because the artist was creative or because he understood how the 'human' system works? Yes.

I read a great essay by Auren Hoffman recently on the role of creative thinking future society. His essay proposes that once computers and AI develop sufficiently, computers will take over more systems-thinking tasks, increasing the value of creative thinking to society. This got me thinking about the role of creativity in engineering. After thinking about how creative and systems thinking interact in problem solving, I’ve concluded that by themselves, creativity or systems thinking alone are not sufficient to come up with great solutions to problems. Combining both effectively is how great solutions happen.

Systems Thinking Without Creativity

The other day, my professor was talking to my class about the importance of geometric understanding in solving machine learning problems. He said something along the lines of, “If all you can do is execute the integrals, then I’ll get a computer to do it, and it can do it about a thousand times faster than you.” When it comes down to jobs like mathematical execution, computers are way better than humans. The edge humans have though, is the understanding of problems. Computers simply execute a cleverly crafted string of instructions to solve a problem. Humans can think about things like context, cause and effect, and implications of a solution when problem-solving. Though computers are invaluable tools for systems-thinking, they don’t help with the creative aspects of problem-solving.

Innovation in engineering happens as a result of scarcity of resources. A project with an infinite budget and few constraints (e.g. must weigh less than 0.5 kg, use less than 10 watts of electricity, and fits in your pocket) is almost certainly won’t produce any innovative solutions. When attacking a problem, problem-solvers need to make choices on how to allocate scarce resources. How someone chooses to allocate these resources determines how good the solution is. The clever use of available problem solving tools, like math, physics, or even sociology can save a lot of brute-force problem solving effort, and result in ‘better’ solutions.

The gradient descent algorithm is a great example of this in optimization. Gradient descent is a brute-force algorithm that will sure enough produce a solution to a minimization problem (e.g. minimize the power use, probability of error, etc). However, through the creative application of mathematical tricks like the Taylor approximation, Gauss and Newton were able to produce optimization algorithms that are much faster and more stable than the gradient descent algorithm.

These elegant solutions are the ones we end up using and can only happen when creativity gets applied to systems. Creativity can produce great solutions, but only when we understand the system in which that creativity manifests itself.

Creativity Without Systems Thinking

Without understanding systems, no amount of creativity will solve problems or reach goals without understanding the system that artistic solution lives in. Creativity is most commonly applied as art. Art generally attempts to communicate, expand, or understand some aspect of the human condition. Art might seem like it’s divorced from systems thinking, but I argue that it is deeply integrated with systems thinking. Why? Because some art is better than other art.

A while back, I composed this piece in Microsoft Paint, while waiting for my house’s internet connection to return: Riley Yeakle, 2012, MS Paint

This piece communicates the frustration, anguish, and disillusionment of a young man unable to connect to the internet with his laptop one summery night. As poignant as this piece is, it isn’t nearly as good art as  The Scream by Edvard Munch.  Munch better applies both artistic resources and methods to convey the essence of his piece. Like engineering, art attacks a problem (the human condition) and we evaluate art based on how well it attacks that problem. The problem it attacks is highly complex and ill-defined, making the evaluation process much more difficult, but nevertheless, the quality of art can be evaluated. Applying systems thinking to art helps artists to produce better artistic ‘solutions’ to understanding/communicating/expanding the human condition.

Like an engineer who understands the ‘system’ of physics, by understanding the ‘system’ of humanity, an artist can take that aspect of the human condition and use her artistic resources to communicate it as elegantly and effectively as possible. Also like the engineer, the artist experiences constraint and scarcity in her work. A canvas can only be so large, and only so many sounds can be played at once and still make sense. In order to produce a great painting, a painter needs to understand how different strokes and colors can communicate the tone of a scene. By choosing and applying these techniques well, an artist can better communicate or advance some aspect of the human experience, making her work a ‘better’ piece of art.

The Art of Engineering and Engineering Art

Great solutions, either in art or engineering are a consequence of the synthesis of both creative and systems thinking.

In art, an artist must apply knowledge about the system of the human experience to effectively convey meaning in her piece. In engineering, creative application of science and math make solutions innovative in the face of scarcity. In both situations, when an infinite number of solutions are possible, the engineer and the artist must choose which ones will do the best job attacking their problems. The effective combination of creativity and systems knowledge to problem solving is the art of engineering, which is not too far from engineering great art.