Can someone explain the approach here? I understand that they transform the problem into a gamified 3-D matrix decomposition, but what exactly is the motivatio

Can someone explain the approach here? I understand that they transform the problem into a gamified 3-D matrix decomposition, but what exactly is the motivation for using RL to beat this decomposition game ? Why not just use, for example, an evolution algorithm to find incrementally better decompositions ?