Vincent van Gogh’s — The Starry Night — 1889 — at Google Cultural Institute
Let’s begin with a recent milestone in the art world. We now live in a world where art and science have collided. Computers have begun to mimic human creativity. It’s really amazing (and scary!) to see what has already been done. I found this particular article fascinating.
This was from September 2015!
If you followed the link to that article, you’d find that computers were able to mimic the style of some of the world’s greatist painters (think Van Gogh, Munch, Picasso) to create a relatively simple scene of row houses along a river. It is incredible to think these were done by a computer! And yet, not, given how far we’ve come with computing!
The next level of fascination for me is the human psyche when it comes to man vs. computer. I picture that if we had a young art prodigy in our midst who was able to paint any scene in the style of several of the great masters, we might say that they are nothing short of incredible imbued with talents from God above. But once we find out that a computer was able to do this (a sophisticated artificial neural network to be more accurate) we may tend to pass it off as not so incredible. Amazing yes, but almost an expectation these days… and we may even begin to find its faults, it’s non-human characteristics. Or maybe we won’t even like it precisely because it was done by a computer, because its imperfections are programmed… on purpose. This is fair since understandably we fear the eventual congruence of machine and humans (unless it makes for a great character! link to Data or Blade Runner). I know I do to a certain extent.
And such are the mind games we play when we apply computers (more specifically optimization algortithms) to patient scheduling. The computer can now help us produce the perfect schedule… a thing of beauty. One that meets all of our patient’s needs and our staff’s desires, with just some inputs and the click of a button. Art, meet Science.
And that’s where it all goes wrong… because the moment we “lose” control to the machine is the moment we start finding fault with all that it provides. Sure, we have control of the inputs. Things like which patients run on which shifts. What are the patient’s durations? How long are the treatment durations? Etc.
But do we have control of ALL of the inputs? Are we also plugging in the constraints of patient transportation? The preferred patient ontimes or preferred seating assignment? And do we have control of the methods used to derive the schedule? Does the computer take into consideration the workload on the patient care techs and provide buffer for adverse events? The questions go on…
And the simple answer is… we (in this case, the users) don’t know! Because it’s all hidden away in the black box we call the computer. Instructions were given to it, and it follows those instructions by creating the best possible schedule based on its constraints and its objective function. This gets into the scienc-y part of it all. So let’s skip that. Ultimately we accept the science because it is beyond our capabilities, but the art, **that we can judge!** and quite frankly, we often don’t care for it. Maybe it’s because we can’t really argue with a computer. And getting mad at it doesn’t really do anything either!
So, in the grand argument of patient scheduling being an art or science, I would posit the difference comes down to “taste”. Where the word taste can be substituted with control. If you are OK with giving up some control, than you can appreciate the science of scheduling. But if you have that rare person in your clinic who manages a satisfactory and workable schedule week after week, then you may be in the art camp.
What really may get your goat is that you (nurses and techs working the floor) have lost control either way. You have given up control to either an artist, or a computer. We just think we’re in control to some degree because we can argue with the artist when the schedule doesn’t work well! 🙂
Where do you fall on the art vs. science argument?