Yes. The goal is to find (or approximate) the median without storing all the elements. Instead, it approximates the median by finding the median of randomly selected samples from the elements.