Comment by avilay

What is the advantage of using evolutionary algos in this case over using something like Thompson Sampling or Contextual Multi-Armed Bandit? (http://www.kdd.org/kdd2017/papers/view/an-efficient-bandit-a...)