Deep reinforcement learning became prominent because of the work of combining Q-learning with DL. The combination is known as deep Q-learning or DQN for Deep Q Network. This algorithm has powered some of the cutting edge examples of DRL, when Google DeepMind used it to make classic Atari games better than humans in 2012. There are many implementations of this algorithm, and Google has even patented it. The current consensus is that Google patented such a base algorithm in order to thwart patent trolls striking at little guys or developers building commercial applications with DQN. It is unlikely that Google would exercise this legally or that it would have to since this algorithm is no longer considered state of the art.
Patent trolling is a practice whereby an often less-than-ethical company will patent any and all manner of inventions just for the...
