An accurate picture of where people live can be a matter of life and death.
While it is possible to use a tool like Google Earth to view nearly any location on the globe, the problem is analyzing the imagery at scale. This is where machine learning comes into play.
Facebook identified structures at scale using a process called weakly supervised learning. After training the model using large batches of photos, then checking over the results, Facebook was able to reach a 99.6% labeling accuracy for positive examples.
Source: https://www.visualcapitalist.com/facebook-machine-learning-world-population-map/
