Contributed by: kellyseal on Thursday, October 08 2015 @ 07:02 am
Last modified on
When you swipe right on someone’s profile, you have a good idea of your physical preferences and what kind of person you are attracted to. Maybe you like a certain body type or hair color or height, but after that, it gets a little more complex. Maybe you like a particular person’s smile, or that they seem to have a wicked sense of humor or arty style in their photography.
While swiping is making it easier to go through profiles and accept or reject someone at whim, is it becoming more automatic the longer we swipe? Are we careful in our choices, or do we casually and without much thought say yes or no? Do we want someone (or something) else to do the swiping for us, someone who knows our preferences as well or even better than we do? It would save time and effort, but do we want to disengage from matching entirely?
At least one researcher is asking the question. He thinks we can be consistent enough in our picks that a computer can pick up on what we like and do the swiping for us. And why shouldn't this be an option?
Harm de Vries, a post-doctoral researcher at the Université de Montreal, thinks computers would make excellent choices on our behalf, and set about proving it with a recent study.
According to an article in PC World, De Vries scraped 10,000 photos from Tinder and gave a computer his opinion of 8,000 of them. Then he let the computer determine his likes and dislikes for the remaining 2,000, thinking this sample was reflective and large enough to be pretty accurate. Unfortunately, it managed to be right only 55% of the time. In other words, it wasn’t much better than closing your eyes and swiping arbitrarily.
De Vries decided to test a larger sample, so he pulled almost 500,000 photos from OkCupid. With a greater number of images to work with, the computer achieved a higher success rate – 68% - but still managed to correctly agree with only two out of every three choices made by DeVries.
Still, De Vries is hopeful that computers can assist in the decision-making process. He thinks even services like Twitter can benefit from computer learning, which is still pretty accurate even compared to human learning. "One of my friends who collaborated with me got to learn my preferences and he managed 76 percent accuracy, so even for humans it's pretty hard."
Choosing who we are attracted to is not an exact science – we are all attracted to people that don’t necessarily have our “favored” characteristics, like dark hair or an athletic body, but they can still possess something that we find very compelling.
The point is – do we want to control our choices, or have a computer just do it for us?