You can easily determine if a neighborhood is doing relatively well. But, can you measure the progress it makes? Well, it’s not that easy. So, the researchers at MIT’s Media Lab developed a computer vision system that can determine the rate of progress or decay the given neighborhood is making.
To do so, the researchers trained a machine learning system to compare 1.6 million pairs of photos taken seven years apart. They took those photos from Google Street View and explored signs of change on a pixel-by-pixel, object-by-object basis. Also, they taught the AI to overlook changes that could skew the measurement, such as seasonal changes in the trees and parked vehicles.
The results were quite interesting. A neighborhood’s chances to thrive don’t depend on people’s income level, housing price or cultural demographics. Instead, they depend on the education level of the residents, safety score, and access to key business districts.
“So it’s not an income story — it’s not that there are rich people there, and they happen to be more educated,” says César Hidalgo, the Asahi Broadcasting Corporation Associate Professor of Media Arts and Sciences and senior author on the paper. “It appears to be more of a skill story.”
The most interesting part was the growth rate. As per some theorists, a safe neighborhood should have more improvement than the dangerous one. However, the progress didn’t accelerate according to their theory.
At this moment, this approach isn’t entirely credible. Its safety assessment only lined up 72 percent of the time while reviewed by the people on Amazon’s Mechanical Turk. However, conflicts grew mainly over those areas where there was comparatively little change. For now, it can at least pinpoint improvement, but a refined version of this AI could help city governments to take a decision about a particular neighborhood before it reaches to disaster.