Assignent Solutionnurs

Reinforcement learning systems in AI:
a. To what extent are modern reinforcement learning systems in AI (such as the DeepMind systems for learning to play Atari and Go) capturing processes that are also at work in human cognition?
b. What do you think are the closest points of correspondence to known systems in the mind or brain, and what are the points of biggest difference?
c. Are some of these systems more human-like than others, and why?
d. What could be done to make these systems more human-like, not necessarily under the banner of “reinforcement learning”?