Reinforcement learning is a very generic and widely applicable framework. If we had any way to solve RL problems, we'd use it everywhere. The way convnets are used everywhere in CV nowadays. The fact that RL is nowhere to be seen in the industry is pretty telling...
-
Also, what's really annoying when talking about this is that people constantly conflate RL as a problem formulation with a specific class of methods for solving such problems.
Absolutely. People equate deep RL methods with the learning framework, which muddies the waters considerably. In any case, deep RL is not well suited for any RL problem that isn't very simple (and if it's simple then other methods are going to be competitive, such as GAs)
