the gain and bane of Perfectionist Reinforcement Learning in neural networks

 AI 

Reinforcement learning is a manifestation of Perfection [perfect, perfected -- noun]. Such that at the heart of neural network learning, it is all about perfecting the reinforcement. That is the goal. Perfection. Such that survivalism becomes another manifestation of said perfection/reinforcement. 

So the main flaw of neural networks is the incapacity to incorporate 'imperfection' in its learning. 

So the corrollary question, is there another way to learn that does not involve perfectionism.