Reinforcement Learning, Fast and Slow

Explore a comprehensive review of recent advancements in deep reinforcement learning (RL) and their implications for cognitive science. Delve into cutting-edge techniques that have significantly improved the sample efficiency of deep RL, challenging the notion that it is too slow to model human learning. Examine the fundamental connection between fast RL and slower, incremental forms of learning. Investigate how these AI methods could potentially revolutionize our understanding of psychology and neuroscience. Cover key topics including deep RL, memory-based approaches, meta-learning, and their potential applications in cognitive modeling.

Intro
Deep RL
Memory
Meta
Conclusion